sqlglot.generator
1from __future__ import annotations 2 3import logging 4import typing as t 5 6from sqlglot import exp 7from sqlglot.errors import ErrorLevel, UnsupportedError, concat_messages 8from sqlglot.helper import apply_index_offset, csv, seq_get 9from sqlglot.time import format_time 10from sqlglot.tokens import TokenType 11 12logger = logging.getLogger("sqlglot") 13 14 15class Generator: 16 """ 17 Generator converts a given syntax tree to the corresponding SQL string. 18 19 Args: 20 pretty: Whether or not to format the produced SQL string. 21 Default: False. 22 identify: Determines when an identifier should be quoted. Possible values are: 23 False (default): Never quote, except in cases where it's mandatory by the dialect. 24 True or 'always': Always quote. 25 'safe': Only quote identifiers that are case insensitive. 26 normalize: Whether or not to normalize identifiers to lowercase. 27 Default: False. 28 pad: Determines the pad size in a formatted string. 29 Default: 2. 30 indent: Determines the indentation size in a formatted string. 31 Default: 2. 32 normalize_functions: Whether or not to normalize all function names. Possible values are: 33 "upper" or True (default): Convert names to uppercase. 34 "lower": Convert names to lowercase. 35 False: Disables function name normalization. 36 unsupported_level: Determines the generator's behavior when it encounters unsupported expressions. 37 Default ErrorLevel.WARN. 38 max_unsupported: Maximum number of unsupported messages to include in a raised UnsupportedError. 39 This is only relevant if unsupported_level is ErrorLevel.RAISE. 40 Default: 3 41 leading_comma: Determines whether or not the comma is leading or trailing in select expressions. 42 This is only relevant when generating in pretty mode. 43 Default: False 44 max_text_width: The max number of characters in a segment before creating new lines in pretty mode. 45 The default is on the smaller end because the length only represents a segment and not the true 46 line length. 47 Default: 80 48 comments: Whether or not to preserve comments in the output SQL code. 49 Default: True 50 """ 51 52 TRANSFORMS = { 53 exp.DateAdd: lambda self, e: self.func( 54 "DATE_ADD", e.this, e.expression, exp.Literal.string(e.text("unit")) 55 ), 56 exp.TsOrDsAdd: lambda self, e: self.func( 57 "TS_OR_DS_ADD", e.this, e.expression, exp.Literal.string(e.text("unit")) 58 ), 59 exp.CaseSpecificColumnConstraint: lambda self, e: f"{'NOT ' if e.args.get('not_') else ''}CASESPECIFIC", 60 exp.CharacterSetColumnConstraint: lambda self, e: f"CHARACTER SET {self.sql(e, 'this')}", 61 exp.CharacterSetProperty: lambda self, e: f"{'DEFAULT ' if e.args.get('default') else ''}CHARACTER SET={self.sql(e, 'this')}", 62 exp.CheckColumnConstraint: lambda self, e: f"CHECK ({self.sql(e, 'this')})", 63 exp.CollateColumnConstraint: lambda self, e: f"COLLATE {self.sql(e, 'this')}", 64 exp.CopyGrantsProperty: lambda self, e: "COPY GRANTS", 65 exp.CommentColumnConstraint: lambda self, e: f"COMMENT {self.sql(e, 'this')}", 66 exp.DateFormatColumnConstraint: lambda self, e: f"FORMAT {self.sql(e, 'this')}", 67 exp.DefaultColumnConstraint: lambda self, e: f"DEFAULT {self.sql(e, 'this')}", 68 exp.EncodeColumnConstraint: lambda self, e: f"ENCODE {self.sql(e, 'this')}", 69 exp.ExecuteAsProperty: lambda self, e: self.naked_property(e), 70 exp.ExternalProperty: lambda self, e: "EXTERNAL", 71 exp.InlineLengthColumnConstraint: lambda self, e: f"INLINE LENGTH {self.sql(e, 'this')}", 72 exp.LanguageProperty: lambda self, e: self.naked_property(e), 73 exp.LocationProperty: lambda self, e: self.naked_property(e), 74 exp.LogProperty: lambda self, e: f"{'NO ' if e.args.get('no') else ''}LOG", 75 exp.MaterializedProperty: lambda self, e: "MATERIALIZED", 76 exp.NoPrimaryIndexProperty: lambda self, e: "NO PRIMARY INDEX", 77 exp.OnCommitProperty: lambda self, e: f"ON COMMIT {'DELETE' if e.args.get('delete') else 'PRESERVE'} ROWS", 78 exp.OnUpdateColumnConstraint: lambda self, e: f"ON UPDATE {self.sql(e, 'this')}", 79 exp.PathColumnConstraint: lambda self, e: f"PATH {self.sql(e, 'this')}", 80 exp.ReturnsProperty: lambda self, e: self.naked_property(e), 81 exp.SetProperty: lambda self, e: f"{'MULTI' if e.args.get('multi') else ''}SET", 82 exp.SettingsProperty: lambda self, e: f"SETTINGS{self.seg('')}{(self.expressions(e))}", 83 exp.SqlSecurityProperty: lambda self, e: f"SQL SECURITY {'DEFINER' if e.args.get('definer') else 'INVOKER'}", 84 exp.StabilityProperty: lambda self, e: e.name, 85 exp.TemporaryProperty: lambda self, e: f"TEMPORARY", 86 exp.ToTableProperty: lambda self, e: f"TO {self.sql(e.this)}", 87 exp.TransientProperty: lambda self, e: "TRANSIENT", 88 exp.TitleColumnConstraint: lambda self, e: f"TITLE {self.sql(e, 'this')}", 89 exp.UppercaseColumnConstraint: lambda self, e: f"UPPERCASE", 90 exp.VarMap: lambda self, e: self.func("MAP", e.args["keys"], e.args["values"]), 91 exp.VolatileProperty: lambda self, e: "VOLATILE", 92 exp.WithJournalTableProperty: lambda self, e: f"WITH JOURNAL TABLE={self.sql(e, 'this')}", 93 } 94 95 # Whether or not null ordering is supported in order by 96 NULL_ORDERING_SUPPORTED = True 97 98 # Whether or not locking reads (i.e. SELECT ... FOR UPDATE/SHARE) are supported 99 LOCKING_READS_SUPPORTED = False 100 101 # Always do union distinct or union all 102 EXPLICIT_UNION = False 103 104 # Wrap derived values in parens, usually standard but spark doesn't support it 105 WRAP_DERIVED_VALUES = True 106 107 # Whether or not create function uses an AS before the RETURN 108 CREATE_FUNCTION_RETURN_AS = True 109 110 # Whether or not MERGE ... WHEN MATCHED BY SOURCE is allowed 111 MATCHED_BY_SOURCE = True 112 113 # Whether or not the INTERVAL expression works only with values like '1 day' 114 SINGLE_STRING_INTERVAL = False 115 116 # Whether or not the plural form of date parts like day (i.e. "days") is supported in INTERVALs 117 INTERVAL_ALLOWS_PLURAL_FORM = True 118 119 # Whether or not the TABLESAMPLE clause supports a method name, like BERNOULLI 120 TABLESAMPLE_WITH_METHOD = True 121 122 # Whether or not to treat the number in TABLESAMPLE (50) as a percentage 123 TABLESAMPLE_SIZE_IS_PERCENT = False 124 125 # Whether or not limit and fetch are supported (possible values: "ALL", "LIMIT", "FETCH") 126 LIMIT_FETCH = "ALL" 127 128 # Whether or not a table is allowed to be renamed with a db 129 RENAME_TABLE_WITH_DB = True 130 131 # The separator for grouping sets and rollups 132 GROUPINGS_SEP = "," 133 134 # The string used for creating an index on a table 135 INDEX_ON = "ON" 136 137 # Whether or not join hints should be generated 138 JOIN_HINTS = True 139 140 # Whether or not table hints should be generated 141 TABLE_HINTS = True 142 143 # Whether or not query hints should be generated 144 QUERY_HINTS = True 145 146 # What kind of separator to use for query hints 147 QUERY_HINT_SEP = ", " 148 149 # Whether or not comparing against booleans (e.g. x IS TRUE) is supported 150 IS_BOOL_ALLOWED = True 151 152 # Whether or not to include the "SET" keyword in the "INSERT ... ON DUPLICATE KEY UPDATE" statement 153 DUPLICATE_KEY_UPDATE_WITH_SET = True 154 155 # Whether or not to generate the limit as TOP <value> instead of LIMIT <value> 156 LIMIT_IS_TOP = False 157 158 # Whether or not to generate INSERT INTO ... RETURNING or INSERT INTO RETURNING ... 159 RETURNING_END = True 160 161 # Whether or not to generate the (+) suffix for columns used in old-style join conditions 162 COLUMN_JOIN_MARKS_SUPPORTED = False 163 164 # https://cloud.google.com/bigquery/docs/reference/standard-sql/query-syntax 165 SELECT_KINDS: t.Tuple[str, ...] = ("STRUCT", "VALUE") 166 167 # Whether or not VALUES statements can be used as derived tables. 168 # MySQL 5 and Redshift do not allow this, so when False, it will convert 169 # SELECT * VALUES into SELECT UNION 170 VALUES_AS_TABLE = True 171 172 TYPE_MAPPING = { 173 exp.DataType.Type.NCHAR: "CHAR", 174 exp.DataType.Type.NVARCHAR: "VARCHAR", 175 exp.DataType.Type.MEDIUMTEXT: "TEXT", 176 exp.DataType.Type.LONGTEXT: "TEXT", 177 exp.DataType.Type.MEDIUMBLOB: "BLOB", 178 exp.DataType.Type.LONGBLOB: "BLOB", 179 exp.DataType.Type.INET: "INET", 180 } 181 182 STAR_MAPPING = { 183 "except": "EXCEPT", 184 "replace": "REPLACE", 185 } 186 187 TIME_PART_SINGULARS = { 188 "microseconds": "microsecond", 189 "seconds": "second", 190 "minutes": "minute", 191 "hours": "hour", 192 "days": "day", 193 "weeks": "week", 194 "months": "month", 195 "quarters": "quarter", 196 "years": "year", 197 } 198 199 TOKEN_MAPPING: t.Dict[TokenType, str] = {} 200 201 STRUCT_DELIMITER = ("<", ">") 202 203 PARAMETER_TOKEN = "@" 204 205 PROPERTIES_LOCATION = { 206 exp.AlgorithmProperty: exp.Properties.Location.POST_CREATE, 207 exp.AutoIncrementProperty: exp.Properties.Location.POST_SCHEMA, 208 exp.BlockCompressionProperty: exp.Properties.Location.POST_NAME, 209 exp.CharacterSetProperty: exp.Properties.Location.POST_SCHEMA, 210 exp.ChecksumProperty: exp.Properties.Location.POST_NAME, 211 exp.CollateProperty: exp.Properties.Location.POST_SCHEMA, 212 exp.CopyGrantsProperty: exp.Properties.Location.POST_SCHEMA, 213 exp.Cluster: exp.Properties.Location.POST_SCHEMA, 214 exp.ClusteredByProperty: exp.Properties.Location.POST_SCHEMA, 215 exp.DataBlocksizeProperty: exp.Properties.Location.POST_NAME, 216 exp.DefinerProperty: exp.Properties.Location.POST_CREATE, 217 exp.DictRange: exp.Properties.Location.POST_SCHEMA, 218 exp.DictProperty: exp.Properties.Location.POST_SCHEMA, 219 exp.DistKeyProperty: exp.Properties.Location.POST_SCHEMA, 220 exp.DistStyleProperty: exp.Properties.Location.POST_SCHEMA, 221 exp.EngineProperty: exp.Properties.Location.POST_SCHEMA, 222 exp.ExecuteAsProperty: exp.Properties.Location.POST_SCHEMA, 223 exp.ExternalProperty: exp.Properties.Location.POST_CREATE, 224 exp.FallbackProperty: exp.Properties.Location.POST_NAME, 225 exp.FileFormatProperty: exp.Properties.Location.POST_WITH, 226 exp.FreespaceProperty: exp.Properties.Location.POST_NAME, 227 exp.IsolatedLoadingProperty: exp.Properties.Location.POST_NAME, 228 exp.JournalProperty: exp.Properties.Location.POST_NAME, 229 exp.LanguageProperty: exp.Properties.Location.POST_SCHEMA, 230 exp.LikeProperty: exp.Properties.Location.POST_SCHEMA, 231 exp.LocationProperty: exp.Properties.Location.POST_SCHEMA, 232 exp.LockingProperty: exp.Properties.Location.POST_ALIAS, 233 exp.LogProperty: exp.Properties.Location.POST_NAME, 234 exp.MaterializedProperty: exp.Properties.Location.POST_CREATE, 235 exp.MergeBlockRatioProperty: exp.Properties.Location.POST_NAME, 236 exp.NoPrimaryIndexProperty: exp.Properties.Location.POST_EXPRESSION, 237 exp.OnCommitProperty: exp.Properties.Location.POST_EXPRESSION, 238 exp.Order: exp.Properties.Location.POST_SCHEMA, 239 exp.PartitionedByProperty: exp.Properties.Location.POST_WITH, 240 exp.PrimaryKey: exp.Properties.Location.POST_SCHEMA, 241 exp.Property: exp.Properties.Location.POST_WITH, 242 exp.ReturnsProperty: exp.Properties.Location.POST_SCHEMA, 243 exp.RowFormatProperty: exp.Properties.Location.POST_SCHEMA, 244 exp.RowFormatDelimitedProperty: exp.Properties.Location.POST_SCHEMA, 245 exp.RowFormatSerdeProperty: exp.Properties.Location.POST_SCHEMA, 246 exp.SchemaCommentProperty: exp.Properties.Location.POST_SCHEMA, 247 exp.SerdeProperties: exp.Properties.Location.POST_SCHEMA, 248 exp.Set: exp.Properties.Location.POST_SCHEMA, 249 exp.SettingsProperty: exp.Properties.Location.POST_SCHEMA, 250 exp.SetProperty: exp.Properties.Location.POST_CREATE, 251 exp.SortKeyProperty: exp.Properties.Location.POST_SCHEMA, 252 exp.SqlSecurityProperty: exp.Properties.Location.POST_CREATE, 253 exp.StabilityProperty: exp.Properties.Location.POST_SCHEMA, 254 exp.TemporaryProperty: exp.Properties.Location.POST_CREATE, 255 exp.ToTableProperty: exp.Properties.Location.POST_SCHEMA, 256 exp.TransientProperty: exp.Properties.Location.POST_CREATE, 257 exp.MergeTreeTTL: exp.Properties.Location.POST_SCHEMA, 258 exp.VolatileProperty: exp.Properties.Location.POST_CREATE, 259 exp.WithDataProperty: exp.Properties.Location.POST_EXPRESSION, 260 exp.WithJournalTableProperty: exp.Properties.Location.POST_NAME, 261 } 262 263 # Keywords that can't be used as unquoted identifier names 264 RESERVED_KEYWORDS: t.Set[str] = set() 265 266 # Expressions whose comments are separated from them for better formatting 267 WITH_SEPARATED_COMMENTS: t.Tuple[t.Type[exp.Expression], ...] = ( 268 exp.Drop, 269 exp.From, 270 exp.Select, 271 exp.Where, 272 exp.With, 273 ) 274 275 # Expressions that can remain unwrapped when appearing in the context of an INTERVAL 276 UNWRAPPED_INTERVAL_VALUES: t.Tuple[t.Type[exp.Expression], ...] = ( 277 exp.Column, 278 exp.Literal, 279 exp.Neg, 280 exp.Paren, 281 ) 282 283 SENTINEL_LINE_BREAK = "__SQLGLOT__LB__" 284 285 # Autofilled 286 INVERSE_TIME_MAPPING: t.Dict[str, str] = {} 287 INVERSE_TIME_TRIE: t.Dict = {} 288 INDEX_OFFSET = 0 289 UNNEST_COLUMN_ONLY = False 290 ALIAS_POST_TABLESAMPLE = False 291 IDENTIFIERS_CAN_START_WITH_DIGIT = False 292 STRICT_STRING_CONCAT = False 293 NORMALIZE_FUNCTIONS: bool | str = "upper" 294 NULL_ORDERING = "nulls_are_small" 295 ESCAPE_LINE_BREAK = False 296 297 can_identify: t.Callable[[str, str | bool], bool] 298 299 # Delimiters for quotes, identifiers and the corresponding escape characters 300 QUOTE_START = "'" 301 QUOTE_END = "'" 302 IDENTIFIER_START = '"' 303 IDENTIFIER_END = '"' 304 STRING_ESCAPE = "'" 305 IDENTIFIER_ESCAPE = '"' 306 307 # Delimiters for bit, hex, byte and raw literals 308 BIT_START: t.Optional[str] = None 309 BIT_END: t.Optional[str] = None 310 HEX_START: t.Optional[str] = None 311 HEX_END: t.Optional[str] = None 312 BYTE_START: t.Optional[str] = None 313 BYTE_END: t.Optional[str] = None 314 315 __slots__ = ( 316 "pretty", 317 "identify", 318 "normalize", 319 "pad", 320 "_indent", 321 "normalize_functions", 322 "unsupported_level", 323 "max_unsupported", 324 "leading_comma", 325 "max_text_width", 326 "comments", 327 "unsupported_messages", 328 "_escaped_quote_end", 329 "_escaped_identifier_end", 330 "_cache", 331 ) 332 333 def __init__( 334 self, 335 pretty: t.Optional[bool] = None, 336 identify: str | bool = False, 337 normalize: bool = False, 338 pad: int = 2, 339 indent: int = 2, 340 normalize_functions: t.Optional[str | bool] = None, 341 unsupported_level: ErrorLevel = ErrorLevel.WARN, 342 max_unsupported: int = 3, 343 leading_comma: bool = False, 344 max_text_width: int = 80, 345 comments: bool = True, 346 ): 347 import sqlglot 348 349 self.pretty = pretty if pretty is not None else sqlglot.pretty 350 self.identify = identify 351 self.normalize = normalize 352 self.pad = pad 353 self._indent = indent 354 self.unsupported_level = unsupported_level 355 self.max_unsupported = max_unsupported 356 self.leading_comma = leading_comma 357 self.max_text_width = max_text_width 358 self.comments = comments 359 360 # This is both a Dialect property and a Generator argument, so we prioritize the latter 361 self.normalize_functions = ( 362 self.NORMALIZE_FUNCTIONS if normalize_functions is None else normalize_functions 363 ) 364 365 self.unsupported_messages: t.List[str] = [] 366 self._escaped_quote_end: str = self.STRING_ESCAPE + self.QUOTE_END 367 self._escaped_identifier_end: str = self.IDENTIFIER_ESCAPE + self.IDENTIFIER_END 368 self._cache: t.Optional[t.Dict[int, str]] = None 369 370 def generate( 371 self, 372 expression: t.Optional[exp.Expression], 373 cache: t.Optional[t.Dict[int, str]] = None, 374 ) -> str: 375 """ 376 Generates the SQL string corresponding to the given syntax tree. 377 378 Args: 379 expression: The syntax tree. 380 cache: An optional sql string cache. This leverages the hash of an Expression 381 which can be slow to compute, so only use it if you set _hash on each node. 382 383 Returns: 384 The SQL string corresponding to `expression`. 385 """ 386 if cache is not None: 387 self._cache = cache 388 389 self.unsupported_messages = [] 390 sql = self.sql(expression).strip() 391 self._cache = None 392 393 if self.unsupported_level == ErrorLevel.IGNORE: 394 return sql 395 396 if self.unsupported_level == ErrorLevel.WARN: 397 for msg in self.unsupported_messages: 398 logger.warning(msg) 399 elif self.unsupported_level == ErrorLevel.RAISE and self.unsupported_messages: 400 raise UnsupportedError(concat_messages(self.unsupported_messages, self.max_unsupported)) 401 402 if self.pretty: 403 sql = sql.replace(self.SENTINEL_LINE_BREAK, "\n") 404 return sql 405 406 def unsupported(self, message: str) -> None: 407 if self.unsupported_level == ErrorLevel.IMMEDIATE: 408 raise UnsupportedError(message) 409 self.unsupported_messages.append(message) 410 411 def sep(self, sep: str = " ") -> str: 412 return f"{sep.strip()}\n" if self.pretty else sep 413 414 def seg(self, sql: str, sep: str = " ") -> str: 415 return f"{self.sep(sep)}{sql}" 416 417 def pad_comment(self, comment: str) -> str: 418 comment = " " + comment if comment[0].strip() else comment 419 comment = comment + " " if comment[-1].strip() else comment 420 return comment 421 422 def maybe_comment( 423 self, 424 sql: str, 425 expression: t.Optional[exp.Expression] = None, 426 comments: t.Optional[t.List[str]] = None, 427 ) -> str: 428 comments = ( 429 ((expression and expression.comments) if comments is None else comments) # type: ignore 430 if self.comments 431 else None 432 ) 433 434 if not comments or isinstance(expression, exp.Binary): 435 return sql 436 437 sep = "\n" if self.pretty else " " 438 comments_sql = sep.join( 439 f"/*{self.pad_comment(comment)}*/" for comment in comments if comment 440 ) 441 442 if not comments_sql: 443 return sql 444 445 if isinstance(expression, self.WITH_SEPARATED_COMMENTS): 446 return ( 447 f"{self.sep()}{comments_sql}{sql}" 448 if sql[0].isspace() 449 else f"{comments_sql}{self.sep()}{sql}" 450 ) 451 452 return f"{sql} {comments_sql}" 453 454 def wrap(self, expression: exp.Expression | str) -> str: 455 this_sql = self.indent( 456 self.sql(expression) 457 if isinstance(expression, (exp.Select, exp.Union)) 458 else self.sql(expression, "this"), 459 level=1, 460 pad=0, 461 ) 462 return f"({self.sep('')}{this_sql}{self.seg(')', sep='')}" 463 464 def no_identify(self, func: t.Callable[..., str], *args, **kwargs) -> str: 465 original = self.identify 466 self.identify = False 467 result = func(*args, **kwargs) 468 self.identify = original 469 return result 470 471 def normalize_func(self, name: str) -> str: 472 if self.normalize_functions == "upper" or self.normalize_functions is True: 473 return name.upper() 474 if self.normalize_functions == "lower": 475 return name.lower() 476 return name 477 478 def indent( 479 self, 480 sql: str, 481 level: int = 0, 482 pad: t.Optional[int] = None, 483 skip_first: bool = False, 484 skip_last: bool = False, 485 ) -> str: 486 if not self.pretty: 487 return sql 488 489 pad = self.pad if pad is None else pad 490 lines = sql.split("\n") 491 492 return "\n".join( 493 line 494 if (skip_first and i == 0) or (skip_last and i == len(lines) - 1) 495 else f"{' ' * (level * self._indent + pad)}{line}" 496 for i, line in enumerate(lines) 497 ) 498 499 def sql( 500 self, 501 expression: t.Optional[str | exp.Expression], 502 key: t.Optional[str] = None, 503 comment: bool = True, 504 ) -> str: 505 if not expression: 506 return "" 507 508 if isinstance(expression, str): 509 return expression 510 511 if key: 512 value = expression.args.get(key) 513 if value: 514 return self.sql(value) 515 return "" 516 517 if self._cache is not None: 518 expression_id = hash(expression) 519 520 if expression_id in self._cache: 521 return self._cache[expression_id] 522 523 transform = self.TRANSFORMS.get(expression.__class__) 524 525 if callable(transform): 526 sql = transform(self, expression) 527 elif transform: 528 sql = transform 529 elif isinstance(expression, exp.Expression): 530 exp_handler_name = f"{expression.key}_sql" 531 532 if hasattr(self, exp_handler_name): 533 sql = getattr(self, exp_handler_name)(expression) 534 elif isinstance(expression, exp.Func): 535 sql = self.function_fallback_sql(expression) 536 elif isinstance(expression, exp.Property): 537 sql = self.property_sql(expression) 538 else: 539 raise ValueError(f"Unsupported expression type {expression.__class__.__name__}") 540 else: 541 raise ValueError(f"Expected an Expression. Received {type(expression)}: {expression}") 542 543 sql = self.maybe_comment(sql, expression) if self.comments and comment else sql 544 545 if self._cache is not None: 546 self._cache[expression_id] = sql 547 return sql 548 549 def uncache_sql(self, expression: exp.Uncache) -> str: 550 table = self.sql(expression, "this") 551 exists_sql = " IF EXISTS" if expression.args.get("exists") else "" 552 return f"UNCACHE TABLE{exists_sql} {table}" 553 554 def cache_sql(self, expression: exp.Cache) -> str: 555 lazy = " LAZY" if expression.args.get("lazy") else "" 556 table = self.sql(expression, "this") 557 options = expression.args.get("options") 558 options = f" OPTIONS({self.sql(options[0])} = {self.sql(options[1])})" if options else "" 559 sql = self.sql(expression, "expression") 560 sql = f" AS{self.sep()}{sql}" if sql else "" 561 sql = f"CACHE{lazy} TABLE {table}{options}{sql}" 562 return self.prepend_ctes(expression, sql) 563 564 def characterset_sql(self, expression: exp.CharacterSet) -> str: 565 if isinstance(expression.parent, exp.Cast): 566 return f"CHAR CHARACTER SET {self.sql(expression, 'this')}" 567 default = "DEFAULT " if expression.args.get("default") else "" 568 return f"{default}CHARACTER SET={self.sql(expression, 'this')}" 569 570 def column_sql(self, expression: exp.Column) -> str: 571 join_mark = " (+)" if expression.args.get("join_mark") else "" 572 573 if join_mark and not self.COLUMN_JOIN_MARKS_SUPPORTED: 574 join_mark = "" 575 self.unsupported("Outer join syntax using the (+) operator is not supported.") 576 577 column = ".".join( 578 self.sql(part) 579 for part in ( 580 expression.args.get("catalog"), 581 expression.args.get("db"), 582 expression.args.get("table"), 583 expression.args.get("this"), 584 ) 585 if part 586 ) 587 588 return f"{column}{join_mark}" 589 590 def columnposition_sql(self, expression: exp.ColumnPosition) -> str: 591 this = self.sql(expression, "this") 592 this = f" {this}" if this else "" 593 position = self.sql(expression, "position") 594 return f"{position}{this}" 595 596 def columndef_sql(self, expression: exp.ColumnDef, sep: str = " ") -> str: 597 column = self.sql(expression, "this") 598 kind = self.sql(expression, "kind") 599 constraints = self.expressions(expression, key="constraints", sep=" ", flat=True) 600 exists = "IF NOT EXISTS " if expression.args.get("exists") else "" 601 kind = f"{sep}{kind}" if kind else "" 602 constraints = f" {constraints}" if constraints else "" 603 position = self.sql(expression, "position") 604 position = f" {position}" if position else "" 605 606 return f"{exists}{column}{kind}{constraints}{position}" 607 608 def columnconstraint_sql(self, expression: exp.ColumnConstraint) -> str: 609 this = self.sql(expression, "this") 610 kind_sql = self.sql(expression, "kind").strip() 611 return f"CONSTRAINT {this} {kind_sql}" if this else kind_sql 612 613 def autoincrementcolumnconstraint_sql(self, _) -> str: 614 return self.token_sql(TokenType.AUTO_INCREMENT) 615 616 def compresscolumnconstraint_sql(self, expression: exp.CompressColumnConstraint) -> str: 617 if isinstance(expression.this, list): 618 this = self.wrap(self.expressions(expression, key="this", flat=True)) 619 else: 620 this = self.sql(expression, "this") 621 622 return f"COMPRESS {this}" 623 624 def generatedasidentitycolumnconstraint_sql( 625 self, expression: exp.GeneratedAsIdentityColumnConstraint 626 ) -> str: 627 this = "" 628 if expression.this is not None: 629 on_null = "ON NULL " if expression.args.get("on_null") else "" 630 this = " ALWAYS " if expression.this else f" BY DEFAULT {on_null}" 631 632 start = expression.args.get("start") 633 start = f"START WITH {start}" if start else "" 634 increment = expression.args.get("increment") 635 increment = f" INCREMENT BY {increment}" if increment else "" 636 minvalue = expression.args.get("minvalue") 637 minvalue = f" MINVALUE {minvalue}" if minvalue else "" 638 maxvalue = expression.args.get("maxvalue") 639 maxvalue = f" MAXVALUE {maxvalue}" if maxvalue else "" 640 cycle = expression.args.get("cycle") 641 cycle_sql = "" 642 643 if cycle is not None: 644 cycle_sql = f"{' NO' if not cycle else ''} CYCLE" 645 cycle_sql = cycle_sql.strip() if not start and not increment else cycle_sql 646 647 sequence_opts = "" 648 if start or increment or cycle_sql: 649 sequence_opts = f"{start}{increment}{minvalue}{maxvalue}{cycle_sql}" 650 sequence_opts = f" ({sequence_opts.strip()})" 651 652 expr = self.sql(expression, "expression") 653 expr = f"({expr})" if expr else "IDENTITY" 654 655 return f"GENERATED{this}AS {expr}{sequence_opts}" 656 657 def notnullcolumnconstraint_sql(self, expression: exp.NotNullColumnConstraint) -> str: 658 return f"{'' if expression.args.get('allow_null') else 'NOT '}NULL" 659 660 def primarykeycolumnconstraint_sql(self, expression: exp.PrimaryKeyColumnConstraint) -> str: 661 desc = expression.args.get("desc") 662 if desc is not None: 663 return f"PRIMARY KEY{' DESC' if desc else ' ASC'}" 664 return f"PRIMARY KEY" 665 666 def uniquecolumnconstraint_sql(self, expression: exp.UniqueColumnConstraint) -> str: 667 this = self.sql(expression, "this") 668 this = f" {this}" if this else "" 669 return f"UNIQUE{this}" 670 671 def createable_sql( 672 self, expression: exp.Create, locations: dict[exp.Properties.Location, list[exp.Property]] 673 ) -> str: 674 return self.sql(expression, "this") 675 676 def create_sql(self, expression: exp.Create) -> str: 677 kind = self.sql(expression, "kind").upper() 678 properties = expression.args.get("properties") 679 properties_locs = self.locate_properties(properties) if properties else {} 680 681 this = self.createable_sql(expression, properties_locs) 682 683 properties_sql = "" 684 if properties_locs.get(exp.Properties.Location.POST_SCHEMA) or properties_locs.get( 685 exp.Properties.Location.POST_WITH 686 ): 687 properties_sql = self.sql( 688 exp.Properties( 689 expressions=[ 690 *properties_locs[exp.Properties.Location.POST_SCHEMA], 691 *properties_locs[exp.Properties.Location.POST_WITH], 692 ] 693 ) 694 ) 695 696 begin = " BEGIN" if expression.args.get("begin") else "" 697 expression_sql = self.sql(expression, "expression") 698 if expression_sql: 699 expression_sql = f"{begin}{self.sep()}{expression_sql}" 700 701 if self.CREATE_FUNCTION_RETURN_AS or not isinstance(expression.expression, exp.Return): 702 if properties_locs.get(exp.Properties.Location.POST_ALIAS): 703 postalias_props_sql = self.properties( 704 exp.Properties( 705 expressions=properties_locs[exp.Properties.Location.POST_ALIAS] 706 ), 707 wrapped=False, 708 ) 709 expression_sql = f" AS {postalias_props_sql}{expression_sql}" 710 else: 711 expression_sql = f" AS{expression_sql}" 712 713 postindex_props_sql = "" 714 if properties_locs.get(exp.Properties.Location.POST_INDEX): 715 postindex_props_sql = self.properties( 716 exp.Properties(expressions=properties_locs[exp.Properties.Location.POST_INDEX]), 717 wrapped=False, 718 prefix=" ", 719 ) 720 721 indexes = self.expressions(expression, key="indexes", indent=False, sep=" ") 722 indexes = f" {indexes}" if indexes else "" 723 index_sql = indexes + postindex_props_sql 724 725 replace = " OR REPLACE" if expression.args.get("replace") else "" 726 unique = " UNIQUE" if expression.args.get("unique") else "" 727 728 postcreate_props_sql = "" 729 if properties_locs.get(exp.Properties.Location.POST_CREATE): 730 postcreate_props_sql = self.properties( 731 exp.Properties(expressions=properties_locs[exp.Properties.Location.POST_CREATE]), 732 sep=" ", 733 prefix=" ", 734 wrapped=False, 735 ) 736 737 modifiers = "".join((replace, unique, postcreate_props_sql)) 738 739 postexpression_props_sql = "" 740 if properties_locs.get(exp.Properties.Location.POST_EXPRESSION): 741 postexpression_props_sql = self.properties( 742 exp.Properties( 743 expressions=properties_locs[exp.Properties.Location.POST_EXPRESSION] 744 ), 745 sep=" ", 746 prefix=" ", 747 wrapped=False, 748 ) 749 750 exists_sql = " IF NOT EXISTS" if expression.args.get("exists") else "" 751 no_schema_binding = ( 752 " WITH NO SCHEMA BINDING" if expression.args.get("no_schema_binding") else "" 753 ) 754 755 clone = self.sql(expression, "clone") 756 clone = f" {clone}" if clone else "" 757 758 expression_sql = f"CREATE{modifiers} {kind}{exists_sql} {this}{properties_sql}{expression_sql}{postexpression_props_sql}{index_sql}{no_schema_binding}{clone}" 759 return self.prepend_ctes(expression, expression_sql) 760 761 def clone_sql(self, expression: exp.Clone) -> str: 762 this = self.sql(expression, "this") 763 when = self.sql(expression, "when") 764 765 if when: 766 kind = self.sql(expression, "kind") 767 expr = self.sql(expression, "expression") 768 return f"CLONE {this} {when} ({kind} => {expr})" 769 770 return f"CLONE {this}" 771 772 def describe_sql(self, expression: exp.Describe) -> str: 773 return f"DESCRIBE {self.sql(expression, 'this')}" 774 775 def prepend_ctes(self, expression: exp.Expression, sql: str) -> str: 776 with_ = self.sql(expression, "with") 777 if with_: 778 sql = f"{with_}{self.sep()}{sql}" 779 return sql 780 781 def with_sql(self, expression: exp.With) -> str: 782 sql = self.expressions(expression, flat=True) 783 recursive = "RECURSIVE " if expression.args.get("recursive") else "" 784 785 return f"WITH {recursive}{sql}" 786 787 def cte_sql(self, expression: exp.CTE) -> str: 788 alias = self.sql(expression, "alias") 789 return f"{alias} AS {self.wrap(expression)}" 790 791 def tablealias_sql(self, expression: exp.TableAlias) -> str: 792 alias = self.sql(expression, "this") 793 columns = self.expressions(expression, key="columns", flat=True) 794 columns = f"({columns})" if columns else "" 795 return f"{alias}{columns}" 796 797 def bitstring_sql(self, expression: exp.BitString) -> str: 798 this = self.sql(expression, "this") 799 if self.BIT_START: 800 return f"{self.BIT_START}{this}{self.BIT_END}" 801 return f"{int(this, 2)}" 802 803 def hexstring_sql(self, expression: exp.HexString) -> str: 804 this = self.sql(expression, "this") 805 if self.HEX_START: 806 return f"{self.HEX_START}{this}{self.HEX_END}" 807 return f"{int(this, 16)}" 808 809 def bytestring_sql(self, expression: exp.ByteString) -> str: 810 this = self.sql(expression, "this") 811 if self.BYTE_START: 812 return f"{self.BYTE_START}{this}{self.BYTE_END}" 813 return this 814 815 def rawstring_sql(self, expression: exp.RawString) -> str: 816 string = self.escape_str(expression.this.replace("\\", "\\\\")) 817 return f"{self.QUOTE_START}{string}{self.QUOTE_END}" 818 819 def datatypesize_sql(self, expression: exp.DataTypeSize) -> str: 820 this = self.sql(expression, "this") 821 specifier = self.sql(expression, "expression") 822 specifier = f" {specifier}" if specifier else "" 823 return f"{this}{specifier}" 824 825 def datatype_sql(self, expression: exp.DataType) -> str: 826 type_value = expression.this 827 type_sql = ( 828 self.TYPE_MAPPING.get(type_value, type_value.value) 829 if isinstance(type_value, exp.DataType.Type) 830 else type_value 831 ) 832 nested = "" 833 interior = self.expressions(expression, flat=True) 834 values = "" 835 if interior: 836 if expression.args.get("nested"): 837 nested = f"{self.STRUCT_DELIMITER[0]}{interior}{self.STRUCT_DELIMITER[1]}" 838 if expression.args.get("values") is not None: 839 delimiters = ("[", "]") if type_value == exp.DataType.Type.ARRAY else ("(", ")") 840 values = self.expressions(expression, key="values", flat=True) 841 values = f"{delimiters[0]}{values}{delimiters[1]}" 842 else: 843 nested = f"({interior})" 844 845 return f"{type_sql}{nested}{values}" 846 847 def directory_sql(self, expression: exp.Directory) -> str: 848 local = "LOCAL " if expression.args.get("local") else "" 849 row_format = self.sql(expression, "row_format") 850 row_format = f" {row_format}" if row_format else "" 851 return f"{local}DIRECTORY {self.sql(expression, 'this')}{row_format}" 852 853 def delete_sql(self, expression: exp.Delete) -> str: 854 this = self.sql(expression, "this") 855 this = f" FROM {this}" if this else "" 856 using = self.sql(expression, "using") 857 using = f" USING {using}" if using else "" 858 where = self.sql(expression, "where") 859 returning = self.sql(expression, "returning") 860 limit = self.sql(expression, "limit") 861 tables = self.expressions(expression, key="tables") 862 tables = f" {tables}" if tables else "" 863 if self.RETURNING_END: 864 expression_sql = f"{this}{using}{where}{returning}{limit}" 865 else: 866 expression_sql = f"{returning}{this}{using}{where}{limit}" 867 return self.prepend_ctes(expression, f"DELETE{tables}{expression_sql}") 868 869 def drop_sql(self, expression: exp.Drop) -> str: 870 this = self.sql(expression, "this") 871 kind = expression.args["kind"] 872 exists_sql = " IF EXISTS " if expression.args.get("exists") else " " 873 temporary = " TEMPORARY" if expression.args.get("temporary") else "" 874 materialized = " MATERIALIZED" if expression.args.get("materialized") else "" 875 cascade = " CASCADE" if expression.args.get("cascade") else "" 876 constraints = " CONSTRAINTS" if expression.args.get("constraints") else "" 877 purge = " PURGE" if expression.args.get("purge") else "" 878 return ( 879 f"DROP{temporary}{materialized} {kind}{exists_sql}{this}{cascade}{constraints}{purge}" 880 ) 881 882 def except_sql(self, expression: exp.Except) -> str: 883 return self.prepend_ctes( 884 expression, 885 self.set_operation(expression, self.except_op(expression)), 886 ) 887 888 def except_op(self, expression: exp.Except) -> str: 889 return f"EXCEPT{'' if expression.args.get('distinct') else ' ALL'}" 890 891 def fetch_sql(self, expression: exp.Fetch) -> str: 892 direction = expression.args.get("direction") 893 direction = f" {direction.upper()}" if direction else "" 894 count = expression.args.get("count") 895 count = f" {count}" if count else "" 896 if expression.args.get("percent"): 897 count = f"{count} PERCENT" 898 with_ties_or_only = "WITH TIES" if expression.args.get("with_ties") else "ONLY" 899 return f"{self.seg('FETCH')}{direction}{count} ROWS {with_ties_or_only}" 900 901 def filter_sql(self, expression: exp.Filter) -> str: 902 this = self.sql(expression, "this") 903 where = self.sql(expression, "expression")[1:] # where has a leading space 904 return f"{this} FILTER({where})" 905 906 def hint_sql(self, expression: exp.Hint) -> str: 907 if not self.QUERY_HINTS: 908 self.unsupported("Hints are not supported") 909 return "" 910 911 return f" /*+ {self.expressions(expression, sep=self.QUERY_HINT_SEP).strip()} */" 912 913 def index_sql(self, expression: exp.Index) -> str: 914 unique = "UNIQUE " if expression.args.get("unique") else "" 915 primary = "PRIMARY " if expression.args.get("primary") else "" 916 amp = "AMP " if expression.args.get("amp") else "" 917 name = self.sql(expression, "this") 918 name = f"{name} " if name else "" 919 table = self.sql(expression, "table") 920 table = f"{self.INDEX_ON} {table} " if table else "" 921 using = self.sql(expression, "using") 922 using = f"USING {using} " if using else "" 923 index = "INDEX " if not table else "" 924 columns = self.expressions(expression, key="columns", flat=True) 925 columns = f"({columns})" if columns else "" 926 partition_by = self.expressions(expression, key="partition_by", flat=True) 927 partition_by = f" PARTITION BY {partition_by}" if partition_by else "" 928 return f"{unique}{primary}{amp}{index}{name}{table}{using}{columns}{partition_by}" 929 930 def identifier_sql(self, expression: exp.Identifier) -> str: 931 text = expression.name 932 lower = text.lower() 933 text = lower if self.normalize and not expression.quoted else text 934 text = text.replace(self.IDENTIFIER_END, self._escaped_identifier_end) 935 if ( 936 expression.quoted 937 or self.can_identify(text, self.identify) 938 or lower in self.RESERVED_KEYWORDS 939 or (not self.IDENTIFIERS_CAN_START_WITH_DIGIT and text[:1].isdigit()) 940 ): 941 text = f"{self.IDENTIFIER_START}{text}{self.IDENTIFIER_END}" 942 return text 943 944 def inputoutputformat_sql(self, expression: exp.InputOutputFormat) -> str: 945 input_format = self.sql(expression, "input_format") 946 input_format = f"INPUTFORMAT {input_format}" if input_format else "" 947 output_format = self.sql(expression, "output_format") 948 output_format = f"OUTPUTFORMAT {output_format}" if output_format else "" 949 return self.sep().join((input_format, output_format)) 950 951 def national_sql(self, expression: exp.National, prefix: str = "N") -> str: 952 string = self.sql(exp.Literal.string(expression.name)) 953 return f"{prefix}{string}" 954 955 def partition_sql(self, expression: exp.Partition) -> str: 956 return f"PARTITION({self.expressions(expression, flat=True)})" 957 958 def properties_sql(self, expression: exp.Properties) -> str: 959 root_properties = [] 960 with_properties = [] 961 962 for p in expression.expressions: 963 p_loc = self.PROPERTIES_LOCATION[p.__class__] 964 if p_loc == exp.Properties.Location.POST_WITH: 965 with_properties.append(p) 966 elif p_loc == exp.Properties.Location.POST_SCHEMA: 967 root_properties.append(p) 968 969 return self.root_properties( 970 exp.Properties(expressions=root_properties) 971 ) + self.with_properties(exp.Properties(expressions=with_properties)) 972 973 def root_properties(self, properties: exp.Properties) -> str: 974 if properties.expressions: 975 return self.sep() + self.expressions(properties, indent=False, sep=" ") 976 return "" 977 978 def properties( 979 self, 980 properties: exp.Properties, 981 prefix: str = "", 982 sep: str = ", ", 983 suffix: str = "", 984 wrapped: bool = True, 985 ) -> str: 986 if properties.expressions: 987 expressions = self.expressions(properties, sep=sep, indent=False) 988 expressions = self.wrap(expressions) if wrapped else expressions 989 return f"{prefix}{' ' if prefix and prefix != ' ' else ''}{expressions}{suffix}" 990 return "" 991 992 def with_properties(self, properties: exp.Properties) -> str: 993 return self.properties(properties, prefix=self.seg("WITH")) 994 995 def locate_properties( 996 self, properties: exp.Properties 997 ) -> t.Dict[exp.Properties.Location, list[exp.Property]]: 998 properties_locs: t.Dict[exp.Properties.Location, list[exp.Property]] = { 999 key: [] for key in exp.Properties.Location 1000 } 1001 1002 for p in properties.expressions: 1003 p_loc = self.PROPERTIES_LOCATION[p.__class__] 1004 if p_loc == exp.Properties.Location.POST_NAME: 1005 properties_locs[exp.Properties.Location.POST_NAME].append(p) 1006 elif p_loc == exp.Properties.Location.POST_INDEX: 1007 properties_locs[exp.Properties.Location.POST_INDEX].append(p) 1008 elif p_loc == exp.Properties.Location.POST_SCHEMA: 1009 properties_locs[exp.Properties.Location.POST_SCHEMA].append(p) 1010 elif p_loc == exp.Properties.Location.POST_WITH: 1011 properties_locs[exp.Properties.Location.POST_WITH].append(p) 1012 elif p_loc == exp.Properties.Location.POST_CREATE: 1013 properties_locs[exp.Properties.Location.POST_CREATE].append(p) 1014 elif p_loc == exp.Properties.Location.POST_ALIAS: 1015 properties_locs[exp.Properties.Location.POST_ALIAS].append(p) 1016 elif p_loc == exp.Properties.Location.POST_EXPRESSION: 1017 properties_locs[exp.Properties.Location.POST_EXPRESSION].append(p) 1018 elif p_loc == exp.Properties.Location.UNSUPPORTED: 1019 self.unsupported(f"Unsupported property {p.key}") 1020 1021 return properties_locs 1022 1023 def property_sql(self, expression: exp.Property) -> str: 1024 property_cls = expression.__class__ 1025 if property_cls == exp.Property: 1026 return f"{expression.name}={self.sql(expression, 'value')}" 1027 1028 property_name = exp.Properties.PROPERTY_TO_NAME.get(property_cls) 1029 if not property_name: 1030 self.unsupported(f"Unsupported property {expression.key}") 1031 1032 return f"{property_name}={self.sql(expression, 'this')}" 1033 1034 def likeproperty_sql(self, expression: exp.LikeProperty) -> str: 1035 options = " ".join(f"{e.name} {self.sql(e, 'value')}" for e in expression.expressions) 1036 options = f" {options}" if options else "" 1037 return f"LIKE {self.sql(expression, 'this')}{options}" 1038 1039 def fallbackproperty_sql(self, expression: exp.FallbackProperty) -> str: 1040 no = "NO " if expression.args.get("no") else "" 1041 protection = " PROTECTION" if expression.args.get("protection") else "" 1042 return f"{no}FALLBACK{protection}" 1043 1044 def journalproperty_sql(self, expression: exp.JournalProperty) -> str: 1045 no = "NO " if expression.args.get("no") else "" 1046 local = expression.args.get("local") 1047 local = f"{local} " if local else "" 1048 dual = "DUAL " if expression.args.get("dual") else "" 1049 before = "BEFORE " if expression.args.get("before") else "" 1050 after = "AFTER " if expression.args.get("after") else "" 1051 return f"{no}{local}{dual}{before}{after}JOURNAL" 1052 1053 def freespaceproperty_sql(self, expression: exp.FreespaceProperty) -> str: 1054 freespace = self.sql(expression, "this") 1055 percent = " PERCENT" if expression.args.get("percent") else "" 1056 return f"FREESPACE={freespace}{percent}" 1057 1058 def checksumproperty_sql(self, expression: exp.ChecksumProperty) -> str: 1059 if expression.args.get("default"): 1060 property = "DEFAULT" 1061 elif expression.args.get("on"): 1062 property = "ON" 1063 else: 1064 property = "OFF" 1065 return f"CHECKSUM={property}" 1066 1067 def mergeblockratioproperty_sql(self, expression: exp.MergeBlockRatioProperty) -> str: 1068 if expression.args.get("no"): 1069 return "NO MERGEBLOCKRATIO" 1070 if expression.args.get("default"): 1071 return "DEFAULT MERGEBLOCKRATIO" 1072 1073 percent = " PERCENT" if expression.args.get("percent") else "" 1074 return f"MERGEBLOCKRATIO={self.sql(expression, 'this')}{percent}" 1075 1076 def datablocksizeproperty_sql(self, expression: exp.DataBlocksizeProperty) -> str: 1077 default = expression.args.get("default") 1078 minimum = expression.args.get("minimum") 1079 maximum = expression.args.get("maximum") 1080 if default or minimum or maximum: 1081 if default: 1082 prop = "DEFAULT" 1083 elif minimum: 1084 prop = "MINIMUM" 1085 else: 1086 prop = "MAXIMUM" 1087 return f"{prop} DATABLOCKSIZE" 1088 units = expression.args.get("units") 1089 units = f" {units}" if units else "" 1090 return f"DATABLOCKSIZE={self.sql(expression, 'size')}{units}" 1091 1092 def blockcompressionproperty_sql(self, expression: exp.BlockCompressionProperty) -> str: 1093 autotemp = expression.args.get("autotemp") 1094 always = expression.args.get("always") 1095 default = expression.args.get("default") 1096 manual = expression.args.get("manual") 1097 never = expression.args.get("never") 1098 1099 if autotemp is not None: 1100 prop = f"AUTOTEMP({self.expressions(autotemp)})" 1101 elif always: 1102 prop = "ALWAYS" 1103 elif default: 1104 prop = "DEFAULT" 1105 elif manual: 1106 prop = "MANUAL" 1107 elif never: 1108 prop = "NEVER" 1109 return f"BLOCKCOMPRESSION={prop}" 1110 1111 def isolatedloadingproperty_sql(self, expression: exp.IsolatedLoadingProperty) -> str: 1112 no = expression.args.get("no") 1113 no = " NO" if no else "" 1114 concurrent = expression.args.get("concurrent") 1115 concurrent = " CONCURRENT" if concurrent else "" 1116 1117 for_ = "" 1118 if expression.args.get("for_all"): 1119 for_ = " FOR ALL" 1120 elif expression.args.get("for_insert"): 1121 for_ = " FOR INSERT" 1122 elif expression.args.get("for_none"): 1123 for_ = " FOR NONE" 1124 return f"WITH{no}{concurrent} ISOLATED LOADING{for_}" 1125 1126 def lockingproperty_sql(self, expression: exp.LockingProperty) -> str: 1127 kind = expression.args.get("kind") 1128 this = f" {self.sql(expression, 'this')}" if expression.this else "" 1129 for_or_in = expression.args.get("for_or_in") 1130 lock_type = expression.args.get("lock_type") 1131 override = " OVERRIDE" if expression.args.get("override") else "" 1132 return f"LOCKING {kind}{this} {for_or_in} {lock_type}{override}" 1133 1134 def withdataproperty_sql(self, expression: exp.WithDataProperty) -> str: 1135 data_sql = f"WITH {'NO ' if expression.args.get('no') else ''}DATA" 1136 statistics = expression.args.get("statistics") 1137 statistics_sql = "" 1138 if statistics is not None: 1139 statistics_sql = f" AND {'NO ' if not statistics else ''}STATISTICS" 1140 return f"{data_sql}{statistics_sql}" 1141 1142 def insert_sql(self, expression: exp.Insert) -> str: 1143 overwrite = expression.args.get("overwrite") 1144 1145 if isinstance(expression.this, exp.Directory): 1146 this = " OVERWRITE" if overwrite else " INTO" 1147 else: 1148 this = " OVERWRITE TABLE" if overwrite else " INTO" 1149 1150 alternative = expression.args.get("alternative") 1151 alternative = f" OR {alternative}" if alternative else "" 1152 ignore = " IGNORE" if expression.args.get("ignore") else "" 1153 1154 this = f"{this} {self.sql(expression, 'this')}" 1155 1156 exists = " IF EXISTS" if expression.args.get("exists") else "" 1157 partition_sql = ( 1158 f" {self.sql(expression, 'partition')}" if expression.args.get("partition") else "" 1159 ) 1160 where = self.sql(expression, "where") 1161 where = f"{self.sep()}REPLACE WHERE {where}" if where else "" 1162 expression_sql = f"{self.sep()}{self.sql(expression, 'expression')}" 1163 conflict = self.sql(expression, "conflict") 1164 returning = self.sql(expression, "returning") 1165 1166 if self.RETURNING_END: 1167 expression_sql = f"{expression_sql}{conflict}{returning}" 1168 else: 1169 expression_sql = f"{returning}{expression_sql}{conflict}" 1170 1171 sql = f"INSERT{alternative}{ignore}{this}{exists}{partition_sql}{where}{expression_sql}" 1172 return self.prepend_ctes(expression, sql) 1173 1174 def intersect_sql(self, expression: exp.Intersect) -> str: 1175 return self.prepend_ctes( 1176 expression, 1177 self.set_operation(expression, self.intersect_op(expression)), 1178 ) 1179 1180 def intersect_op(self, expression: exp.Intersect) -> str: 1181 return f"INTERSECT{'' if expression.args.get('distinct') else ' ALL'}" 1182 1183 def introducer_sql(self, expression: exp.Introducer) -> str: 1184 return f"{self.sql(expression, 'this')} {self.sql(expression, 'expression')}" 1185 1186 def pseudotype_sql(self, expression: exp.PseudoType) -> str: 1187 return expression.name.upper() 1188 1189 def onconflict_sql(self, expression: exp.OnConflict) -> str: 1190 conflict = "ON DUPLICATE KEY" if expression.args.get("duplicate") else "ON CONFLICT" 1191 constraint = self.sql(expression, "constraint") 1192 if constraint: 1193 constraint = f"ON CONSTRAINT {constraint}" 1194 key = self.expressions(expression, key="key", flat=True) 1195 do = "" if expression.args.get("duplicate") else " DO " 1196 nothing = "NOTHING" if expression.args.get("nothing") else "" 1197 expressions = self.expressions(expression, flat=True) 1198 set_keyword = "SET " if self.DUPLICATE_KEY_UPDATE_WITH_SET else "" 1199 if expressions: 1200 expressions = f"UPDATE {set_keyword}{expressions}" 1201 return f"{self.seg(conflict)} {constraint}{key}{do}{nothing}{expressions}" 1202 1203 def returning_sql(self, expression: exp.Returning) -> str: 1204 return f"{self.seg('RETURNING')} {self.expressions(expression, flat=True)}" 1205 1206 def rowformatdelimitedproperty_sql(self, expression: exp.RowFormatDelimitedProperty) -> str: 1207 fields = expression.args.get("fields") 1208 fields = f" FIELDS TERMINATED BY {fields}" if fields else "" 1209 escaped = expression.args.get("escaped") 1210 escaped = f" ESCAPED BY {escaped}" if escaped else "" 1211 items = expression.args.get("collection_items") 1212 items = f" COLLECTION ITEMS TERMINATED BY {items}" if items else "" 1213 keys = expression.args.get("map_keys") 1214 keys = f" MAP KEYS TERMINATED BY {keys}" if keys else "" 1215 lines = expression.args.get("lines") 1216 lines = f" LINES TERMINATED BY {lines}" if lines else "" 1217 null = expression.args.get("null") 1218 null = f" NULL DEFINED AS {null}" if null else "" 1219 return f"ROW FORMAT DELIMITED{fields}{escaped}{items}{keys}{lines}{null}" 1220 1221 def withtablehint_sql(self, expression: exp.WithTableHint) -> str: 1222 return f"WITH ({self.expressions(expression, flat=True)})" 1223 1224 def indextablehint_sql(self, expression: exp.IndexTableHint) -> str: 1225 this = f"{self.sql(expression, 'this')} INDEX" 1226 target = self.sql(expression, "target") 1227 target = f" FOR {target}" if target else "" 1228 return f"{this}{target} ({self.expressions(expression, flat=True)})" 1229 1230 def table_sql(self, expression: exp.Table, sep: str = " AS ") -> str: 1231 table = ".".join( 1232 part 1233 for part in [ 1234 self.sql(expression, "catalog"), 1235 self.sql(expression, "db"), 1236 self.sql(expression, "this"), 1237 ] 1238 if part 1239 ) 1240 1241 alias = self.sql(expression, "alias") 1242 alias = f"{sep}{alias}" if alias else "" 1243 hints = self.expressions(expression, key="hints", sep=" ") 1244 hints = f" {hints}" if hints and self.TABLE_HINTS else "" 1245 pivots = self.expressions(expression, key="pivots", sep=" ", flat=True) 1246 pivots = f" {pivots}" if pivots else "" 1247 joins = self.expressions(expression, key="joins", sep="", skip_first=True) 1248 laterals = self.expressions(expression, key="laterals", sep="") 1249 system_time = expression.args.get("system_time") 1250 system_time = f" {self.sql(expression, 'system_time')}" if system_time else "" 1251 1252 return f"{table}{system_time}{alias}{hints}{pivots}{joins}{laterals}" 1253 1254 def tablesample_sql( 1255 self, expression: exp.TableSample, seed_prefix: str = "SEED", sep=" AS " 1256 ) -> str: 1257 if self.ALIAS_POST_TABLESAMPLE and expression.this.alias: 1258 table = expression.this.copy() 1259 table.set("alias", None) 1260 this = self.sql(table) 1261 alias = f"{sep}{self.sql(expression.this, 'alias')}" 1262 else: 1263 this = self.sql(expression, "this") 1264 alias = "" 1265 method = self.sql(expression, "method") 1266 method = f"{method.upper()} " if method and self.TABLESAMPLE_WITH_METHOD else "" 1267 numerator = self.sql(expression, "bucket_numerator") 1268 denominator = self.sql(expression, "bucket_denominator") 1269 field = self.sql(expression, "bucket_field") 1270 field = f" ON {field}" if field else "" 1271 bucket = f"BUCKET {numerator} OUT OF {denominator}{field}" if numerator else "" 1272 percent = self.sql(expression, "percent") 1273 percent = f"{percent} PERCENT" if percent else "" 1274 rows = self.sql(expression, "rows") 1275 rows = f"{rows} ROWS" if rows else "" 1276 size = self.sql(expression, "size") 1277 if size and self.TABLESAMPLE_SIZE_IS_PERCENT: 1278 size = f"{size} PERCENT" 1279 seed = self.sql(expression, "seed") 1280 seed = f" {seed_prefix} ({seed})" if seed else "" 1281 kind = expression.args.get("kind", "TABLESAMPLE") 1282 return f"{this} {kind} {method}({bucket}{percent}{rows}{size}){seed}{alias}" 1283 1284 def pivot_sql(self, expression: exp.Pivot) -> str: 1285 expressions = self.expressions(expression, flat=True) 1286 1287 if expression.this: 1288 this = self.sql(expression, "this") 1289 on = f"{self.seg('ON')} {expressions}" 1290 using = self.expressions(expression, key="using", flat=True) 1291 using = f"{self.seg('USING')} {using}" if using else "" 1292 group = self.sql(expression, "group") 1293 return f"PIVOT {this}{on}{using}{group}" 1294 1295 alias = self.sql(expression, "alias") 1296 alias = f" AS {alias}" if alias else "" 1297 unpivot = expression.args.get("unpivot") 1298 direction = "UNPIVOT" if unpivot else "PIVOT" 1299 field = self.sql(expression, "field") 1300 return f"{direction}({expressions} FOR {field}){alias}" 1301 1302 def tuple_sql(self, expression: exp.Tuple) -> str: 1303 return f"({self.expressions(expression, flat=True)})" 1304 1305 def update_sql(self, expression: exp.Update) -> str: 1306 this = self.sql(expression, "this") 1307 set_sql = self.expressions(expression, flat=True) 1308 from_sql = self.sql(expression, "from") 1309 where_sql = self.sql(expression, "where") 1310 returning = self.sql(expression, "returning") 1311 limit = self.sql(expression, "limit") 1312 if self.RETURNING_END: 1313 expression_sql = f"{from_sql}{where_sql}{returning}{limit}" 1314 else: 1315 expression_sql = f"{returning}{from_sql}{where_sql}{limit}" 1316 sql = f"UPDATE {this} SET {set_sql}{expression_sql}" 1317 return self.prepend_ctes(expression, sql) 1318 1319 def values_sql(self, expression: exp.Values) -> str: 1320 # The VALUES clause is still valid in an `INSERT INTO ..` statement, for example 1321 if self.VALUES_AS_TABLE or not expression.find_ancestor(exp.From, exp.Join): 1322 args = self.expressions(expression) 1323 alias = self.sql(expression, "alias") 1324 values = f"VALUES{self.seg('')}{args}" 1325 values = ( 1326 f"({values})" 1327 if self.WRAP_DERIVED_VALUES and (alias or isinstance(expression.parent, exp.From)) 1328 else values 1329 ) 1330 return f"{values} AS {alias}" if alias else values 1331 1332 # Converts `VALUES...` expression into a series of select unions. 1333 # Note: If you have a lot of unions then this will result in a large number of recursive statements to 1334 # evaluate the expression. You may need to increase `sys.setrecursionlimit` to run and it can also be 1335 # very slow. 1336 expression = expression.copy() 1337 column_names = expression.alias and expression.args["alias"].columns 1338 1339 selects = [] 1340 1341 for i, tup in enumerate(expression.expressions): 1342 row = tup.expressions 1343 1344 if i == 0 and column_names: 1345 row = [ 1346 exp.alias_(value, column_name) for value, column_name in zip(row, column_names) 1347 ] 1348 1349 selects.append(exp.Select(expressions=row)) 1350 1351 subquery_expression: exp.Select | exp.Union = selects[0] 1352 if len(selects) > 1: 1353 for select in selects[1:]: 1354 subquery_expression = exp.union( 1355 subquery_expression, select, distinct=False, copy=False 1356 ) 1357 1358 return self.subquery_sql(subquery_expression.subquery(expression.alias, copy=False)) 1359 1360 def var_sql(self, expression: exp.Var) -> str: 1361 return self.sql(expression, "this") 1362 1363 def into_sql(self, expression: exp.Into) -> str: 1364 temporary = " TEMPORARY" if expression.args.get("temporary") else "" 1365 unlogged = " UNLOGGED" if expression.args.get("unlogged") else "" 1366 return f"{self.seg('INTO')}{temporary or unlogged} {self.sql(expression, 'this')}" 1367 1368 def from_sql(self, expression: exp.From) -> str: 1369 return f"{self.seg('FROM')} {self.sql(expression, 'this')}" 1370 1371 def group_sql(self, expression: exp.Group) -> str: 1372 group_by = self.op_expressions("GROUP BY", expression) 1373 1374 if expression.args.get("all"): 1375 return f"{group_by} ALL" 1376 1377 grouping_sets = self.expressions(expression, key="grouping_sets", indent=False) 1378 grouping_sets = ( 1379 f"{self.seg('GROUPING SETS')} {self.wrap(grouping_sets)}" if grouping_sets else "" 1380 ) 1381 1382 cube = expression.args.get("cube", []) 1383 if seq_get(cube, 0) is True: 1384 return f"{group_by}{self.seg('WITH CUBE')}" 1385 else: 1386 cube_sql = self.expressions(expression, key="cube", indent=False) 1387 cube_sql = f"{self.seg('CUBE')} {self.wrap(cube_sql)}" if cube_sql else "" 1388 1389 rollup = expression.args.get("rollup", []) 1390 if seq_get(rollup, 0) is True: 1391 return f"{group_by}{self.seg('WITH ROLLUP')}" 1392 else: 1393 rollup_sql = self.expressions(expression, key="rollup", indent=False) 1394 rollup_sql = f"{self.seg('ROLLUP')} {self.wrap(rollup_sql)}" if rollup_sql else "" 1395 1396 groupings = csv( 1397 grouping_sets, 1398 cube_sql, 1399 rollup_sql, 1400 self.seg("WITH TOTALS") if expression.args.get("totals") else "", 1401 sep=self.GROUPINGS_SEP, 1402 ) 1403 1404 if expression.args.get("expressions") and groupings: 1405 group_by = f"{group_by}{self.GROUPINGS_SEP}" 1406 1407 return f"{group_by}{groupings}" 1408 1409 def having_sql(self, expression: exp.Having) -> str: 1410 this = self.indent(self.sql(expression, "this")) 1411 return f"{self.seg('HAVING')}{self.sep()}{this}" 1412 1413 def join_sql(self, expression: exp.Join) -> str: 1414 op_sql = " ".join( 1415 op 1416 for op in ( 1417 expression.method, 1418 "GLOBAL" if expression.args.get("global") else None, 1419 expression.side, 1420 expression.kind, 1421 expression.hint if self.JOIN_HINTS else None, 1422 ) 1423 if op 1424 ) 1425 on_sql = self.sql(expression, "on") 1426 using = expression.args.get("using") 1427 1428 if not on_sql and using: 1429 on_sql = csv(*(self.sql(column) for column in using)) 1430 1431 this_sql = self.sql(expression, "this") 1432 1433 if on_sql: 1434 on_sql = self.indent(on_sql, skip_first=True) 1435 space = self.seg(" " * self.pad) if self.pretty else " " 1436 if using: 1437 on_sql = f"{space}USING ({on_sql})" 1438 else: 1439 on_sql = f"{space}ON {on_sql}" 1440 elif not op_sql: 1441 return f", {this_sql}" 1442 1443 op_sql = f"{op_sql} JOIN" if op_sql else "JOIN" 1444 return f"{self.seg(op_sql)} {this_sql}{on_sql}" 1445 1446 def lambda_sql(self, expression: exp.Lambda, arrow_sep: str = "->") -> str: 1447 args = self.expressions(expression, flat=True) 1448 args = f"({args})" if len(args.split(",")) > 1 else args 1449 return f"{args} {arrow_sep} {self.sql(expression, 'this')}" 1450 1451 def lateral_sql(self, expression: exp.Lateral) -> str: 1452 this = self.sql(expression, "this") 1453 1454 if isinstance(expression.this, exp.Subquery): 1455 return f"LATERAL {this}" 1456 1457 if expression.args.get("view"): 1458 alias = expression.args["alias"] 1459 columns = self.expressions(alias, key="columns", flat=True) 1460 table = f" {alias.name}" if alias.name else "" 1461 columns = f" AS {columns}" if columns else "" 1462 op_sql = self.seg(f"LATERAL VIEW{' OUTER' if expression.args.get('outer') else ''}") 1463 return f"{op_sql}{self.sep()}{this}{table}{columns}" 1464 1465 alias = self.sql(expression, "alias") 1466 alias = f" AS {alias}" if alias else "" 1467 return f"LATERAL {this}{alias}" 1468 1469 def limit_sql(self, expression: exp.Limit, top: bool = False) -> str: 1470 this = self.sql(expression, "this") 1471 args = ", ".join( 1472 sql 1473 for sql in ( 1474 self.sql(expression, "offset"), 1475 self.sql(expression, "expression"), 1476 ) 1477 if sql 1478 ) 1479 return f"{this}{self.seg('TOP' if top else 'LIMIT')} {args}" 1480 1481 def offset_sql(self, expression: exp.Offset) -> str: 1482 this = self.sql(expression, "this") 1483 return f"{this}{self.seg('OFFSET')} {self.sql(expression, 'expression')}" 1484 1485 def setitem_sql(self, expression: exp.SetItem) -> str: 1486 kind = self.sql(expression, "kind") 1487 kind = f"{kind} " if kind else "" 1488 this = self.sql(expression, "this") 1489 expressions = self.expressions(expression) 1490 collate = self.sql(expression, "collate") 1491 collate = f" COLLATE {collate}" if collate else "" 1492 global_ = "GLOBAL " if expression.args.get("global") else "" 1493 return f"{global_}{kind}{this}{expressions}{collate}" 1494 1495 def set_sql(self, expression: exp.Set) -> str: 1496 expressions = ( 1497 f" {self.expressions(expression, flat=True)}" if expression.expressions else "" 1498 ) 1499 tag = " TAG" if expression.args.get("tag") else "" 1500 return f"{'UNSET' if expression.args.get('unset') else 'SET'}{tag}{expressions}" 1501 1502 def pragma_sql(self, expression: exp.Pragma) -> str: 1503 return f"PRAGMA {self.sql(expression, 'this')}" 1504 1505 def lock_sql(self, expression: exp.Lock) -> str: 1506 if not self.LOCKING_READS_SUPPORTED: 1507 self.unsupported("Locking reads using 'FOR UPDATE/SHARE' are not supported") 1508 return "" 1509 1510 lock_type = "FOR UPDATE" if expression.args["update"] else "FOR SHARE" 1511 expressions = self.expressions(expression, flat=True) 1512 expressions = f" OF {expressions}" if expressions else "" 1513 wait = expression.args.get("wait") 1514 1515 if wait is not None: 1516 if isinstance(wait, exp.Literal): 1517 wait = f" WAIT {self.sql(wait)}" 1518 else: 1519 wait = " NOWAIT" if wait else " SKIP LOCKED" 1520 1521 return f"{lock_type}{expressions}{wait or ''}" 1522 1523 def literal_sql(self, expression: exp.Literal) -> str: 1524 text = expression.this or "" 1525 if expression.is_string: 1526 text = f"{self.QUOTE_START}{self.escape_str(text)}{self.QUOTE_END}" 1527 return text 1528 1529 def escape_str(self, text: str) -> str: 1530 text = text.replace(self.QUOTE_END, self._escaped_quote_end) 1531 if self.ESCAPE_LINE_BREAK: 1532 text = text.replace("\n", "\\n") 1533 elif self.pretty: 1534 text = text.replace("\n", self.SENTINEL_LINE_BREAK) 1535 return text 1536 1537 def loaddata_sql(self, expression: exp.LoadData) -> str: 1538 local = " LOCAL" if expression.args.get("local") else "" 1539 inpath = f" INPATH {self.sql(expression, 'inpath')}" 1540 overwrite = " OVERWRITE" if expression.args.get("overwrite") else "" 1541 this = f" INTO TABLE {self.sql(expression, 'this')}" 1542 partition = self.sql(expression, "partition") 1543 partition = f" {partition}" if partition else "" 1544 input_format = self.sql(expression, "input_format") 1545 input_format = f" INPUTFORMAT {input_format}" if input_format else "" 1546 serde = self.sql(expression, "serde") 1547 serde = f" SERDE {serde}" if serde else "" 1548 return f"LOAD DATA{local}{inpath}{overwrite}{this}{partition}{input_format}{serde}" 1549 1550 def null_sql(self, *_) -> str: 1551 return "NULL" 1552 1553 def boolean_sql(self, expression: exp.Boolean) -> str: 1554 return "TRUE" if expression.this else "FALSE" 1555 1556 def order_sql(self, expression: exp.Order, flat: bool = False) -> str: 1557 this = self.sql(expression, "this") 1558 this = f"{this} " if this else this 1559 return self.op_expressions(f"{this}ORDER BY", expression, flat=this or flat) # type: ignore 1560 1561 def cluster_sql(self, expression: exp.Cluster) -> str: 1562 return self.op_expressions("CLUSTER BY", expression) 1563 1564 def distribute_sql(self, expression: exp.Distribute) -> str: 1565 return self.op_expressions("DISTRIBUTE BY", expression) 1566 1567 def sort_sql(self, expression: exp.Sort) -> str: 1568 return self.op_expressions("SORT BY", expression) 1569 1570 def ordered_sql(self, expression: exp.Ordered) -> str: 1571 desc = expression.args.get("desc") 1572 asc = not desc 1573 1574 nulls_first = expression.args.get("nulls_first") 1575 nulls_last = not nulls_first 1576 nulls_are_large = self.NULL_ORDERING == "nulls_are_large" 1577 nulls_are_small = self.NULL_ORDERING == "nulls_are_small" 1578 nulls_are_last = self.NULL_ORDERING == "nulls_are_last" 1579 1580 sort_order = " DESC" if desc else "" 1581 nulls_sort_change = "" 1582 if nulls_first and ( 1583 (asc and nulls_are_large) or (desc and nulls_are_small) or nulls_are_last 1584 ): 1585 nulls_sort_change = " NULLS FIRST" 1586 elif ( 1587 nulls_last 1588 and ((asc and nulls_are_small) or (desc and nulls_are_large)) 1589 and not nulls_are_last 1590 ): 1591 nulls_sort_change = " NULLS LAST" 1592 1593 if nulls_sort_change and not self.NULL_ORDERING_SUPPORTED: 1594 self.unsupported( 1595 "Sorting in an ORDER BY on NULLS FIRST/NULLS LAST is not supported by this dialect" 1596 ) 1597 nulls_sort_change = "" 1598 1599 return f"{self.sql(expression, 'this')}{sort_order}{nulls_sort_change}" 1600 1601 def matchrecognize_sql(self, expression: exp.MatchRecognize) -> str: 1602 partition = self.partition_by_sql(expression) 1603 order = self.sql(expression, "order") 1604 measures = self.expressions(expression, key="measures") 1605 measures = self.seg(f"MEASURES{self.seg(measures)}") if measures else "" 1606 rows = self.sql(expression, "rows") 1607 rows = self.seg(rows) if rows else "" 1608 after = self.sql(expression, "after") 1609 after = self.seg(after) if after else "" 1610 pattern = self.sql(expression, "pattern") 1611 pattern = self.seg(f"PATTERN ({pattern})") if pattern else "" 1612 definition_sqls = [ 1613 f"{self.sql(definition, 'alias')} AS {self.sql(definition, 'this')}" 1614 for definition in expression.args.get("define", []) 1615 ] 1616 definitions = self.expressions(sqls=definition_sqls) 1617 define = self.seg(f"DEFINE{self.seg(definitions)}") if definitions else "" 1618 body = "".join( 1619 ( 1620 partition, 1621 order, 1622 measures, 1623 rows, 1624 after, 1625 pattern, 1626 define, 1627 ) 1628 ) 1629 alias = self.sql(expression, "alias") 1630 alias = f" {alias}" if alias else "" 1631 return f"{self.seg('MATCH_RECOGNIZE')} {self.wrap(body)}{alias}" 1632 1633 def query_modifiers(self, expression: exp.Expression, *sqls: str) -> str: 1634 limit: t.Optional[exp.Fetch | exp.Limit] = expression.args.get("limit") 1635 1636 # If the limit is generated as TOP, we need to ensure it's not generated twice 1637 with_offset_limit_modifiers = not isinstance(limit, exp.Limit) or not self.LIMIT_IS_TOP 1638 1639 if self.LIMIT_FETCH == "LIMIT" and isinstance(limit, exp.Fetch): 1640 limit = exp.Limit(expression=limit.args.get("count")) 1641 elif self.LIMIT_FETCH == "FETCH" and isinstance(limit, exp.Limit): 1642 limit = exp.Fetch(direction="FIRST", count=limit.expression) 1643 1644 fetch = isinstance(limit, exp.Fetch) 1645 1646 offset_limit_modifiers = ( 1647 self.offset_limit_modifiers(expression, fetch, limit) 1648 if with_offset_limit_modifiers 1649 else [] 1650 ) 1651 1652 return csv( 1653 *sqls, 1654 *[self.sql(join) for join in expression.args.get("joins") or []], 1655 self.sql(expression, "match"), 1656 *[self.sql(lateral) for lateral in expression.args.get("laterals") or []], 1657 self.sql(expression, "where"), 1658 self.sql(expression, "group"), 1659 self.sql(expression, "having"), 1660 *self.after_having_modifiers(expression), 1661 self.sql(expression, "order"), 1662 *offset_limit_modifiers, 1663 *self.after_limit_modifiers(expression), 1664 sep="", 1665 ) 1666 1667 def offset_limit_modifiers( 1668 self, expression: exp.Expression, fetch: bool, limit: t.Optional[exp.Fetch | exp.Limit] 1669 ) -> t.List[str]: 1670 return [ 1671 self.sql(expression, "offset") if fetch else self.sql(limit), 1672 self.sql(limit) if fetch else self.sql(expression, "offset"), 1673 ] 1674 1675 def after_having_modifiers(self, expression: exp.Expression) -> t.List[str]: 1676 return [ 1677 self.sql(expression, "qualify"), 1678 self.seg("WINDOW ") + self.expressions(expression, key="windows", flat=True) 1679 if expression.args.get("windows") 1680 else "", 1681 self.sql(expression, "distribute"), 1682 self.sql(expression, "sort"), 1683 self.sql(expression, "cluster"), 1684 ] 1685 1686 def after_limit_modifiers(self, expression: exp.Expression) -> t.List[str]: 1687 locks = self.expressions(expression, key="locks", sep=" ") 1688 locks = f" {locks}" if locks else "" 1689 return [locks, self.sql(expression, "sample")] 1690 1691 def select_sql(self, expression: exp.Select) -> str: 1692 hint = self.sql(expression, "hint") 1693 distinct = self.sql(expression, "distinct") 1694 distinct = f" {distinct}" if distinct else "" 1695 kind = self.sql(expression, "kind").upper() 1696 limit = expression.args.get("limit") 1697 top = ( 1698 self.limit_sql(limit, top=True) 1699 if isinstance(limit, exp.Limit) and self.LIMIT_IS_TOP 1700 else "" 1701 ) 1702 1703 expressions = self.expressions(expression) 1704 1705 if kind: 1706 if kind in self.SELECT_KINDS: 1707 kind = f" AS {kind}" 1708 else: 1709 if kind == "STRUCT": 1710 expressions = self.expressions( 1711 sqls=[ 1712 self.sql( 1713 exp.Struct( 1714 expressions=[ 1715 exp.column(e.output_name).eq( 1716 e.this if isinstance(e, exp.Alias) else e 1717 ) 1718 for e in expression.expressions 1719 ] 1720 ) 1721 ) 1722 ] 1723 ) 1724 kind = "" 1725 1726 expressions = f"{self.sep()}{expressions}" if expressions else expressions 1727 sql = self.query_modifiers( 1728 expression, 1729 f"SELECT{top}{hint}{distinct}{kind}{expressions}", 1730 self.sql(expression, "into", comment=False), 1731 self.sql(expression, "from", comment=False), 1732 ) 1733 return self.prepend_ctes(expression, sql) 1734 1735 def schema_sql(self, expression: exp.Schema) -> str: 1736 this = self.sql(expression, "this") 1737 this = f"{this} " if this else "" 1738 sql = self.schema_columns_sql(expression) 1739 return f"{this}{sql}" 1740 1741 def schema_columns_sql(self, expression: exp.Schema) -> str: 1742 return f"({self.sep('')}{self.expressions(expression)}{self.seg(')', sep='')}" 1743 1744 def star_sql(self, expression: exp.Star) -> str: 1745 except_ = self.expressions(expression, key="except", flat=True) 1746 except_ = f"{self.seg(self.STAR_MAPPING['except'])} ({except_})" if except_ else "" 1747 replace = self.expressions(expression, key="replace", flat=True) 1748 replace = f"{self.seg(self.STAR_MAPPING['replace'])} ({replace})" if replace else "" 1749 return f"*{except_}{replace}" 1750 1751 def parameter_sql(self, expression: exp.Parameter) -> str: 1752 this = self.sql(expression, "this") 1753 this = f"{{{this}}}" if expression.args.get("wrapped") else f"{this}" 1754 return f"{self.PARAMETER_TOKEN}{this}" 1755 1756 def sessionparameter_sql(self, expression: exp.SessionParameter) -> str: 1757 this = self.sql(expression, "this") 1758 kind = expression.text("kind") 1759 if kind: 1760 kind = f"{kind}." 1761 return f"@@{kind}{this}" 1762 1763 def placeholder_sql(self, expression: exp.Placeholder) -> str: 1764 return f":{expression.name}" if expression.name else "?" 1765 1766 def subquery_sql(self, expression: exp.Subquery, sep: str = " AS ") -> str: 1767 alias = self.sql(expression, "alias") 1768 alias = f"{sep}{alias}" if alias else "" 1769 1770 pivots = self.expressions(expression, key="pivots", sep=" ", flat=True) 1771 pivots = f" {pivots}" if pivots else "" 1772 1773 sql = self.query_modifiers(expression, self.wrap(expression), alias, pivots) 1774 return self.prepend_ctes(expression, sql) 1775 1776 def qualify_sql(self, expression: exp.Qualify) -> str: 1777 this = self.indent(self.sql(expression, "this")) 1778 return f"{self.seg('QUALIFY')}{self.sep()}{this}" 1779 1780 def union_sql(self, expression: exp.Union) -> str: 1781 return self.prepend_ctes( 1782 expression, 1783 self.set_operation(expression, self.union_op(expression)), 1784 ) 1785 1786 def union_op(self, expression: exp.Union) -> str: 1787 kind = " DISTINCT" if self.EXPLICIT_UNION else "" 1788 kind = kind if expression.args.get("distinct") else " ALL" 1789 return f"UNION{kind}" 1790 1791 def unnest_sql(self, expression: exp.Unnest) -> str: 1792 args = self.expressions(expression, flat=True) 1793 alias = expression.args.get("alias") 1794 if alias and self.UNNEST_COLUMN_ONLY: 1795 columns = alias.columns 1796 alias = self.sql(columns[0]) if columns else "" 1797 else: 1798 alias = self.sql(expression, "alias") 1799 alias = f" AS {alias}" if alias else alias 1800 ordinality = " WITH ORDINALITY" if expression.args.get("ordinality") else "" 1801 offset = expression.args.get("offset") 1802 offset = f" WITH OFFSET AS {self.sql(offset)}" if offset else "" 1803 return f"UNNEST({args}){ordinality}{alias}{offset}" 1804 1805 def where_sql(self, expression: exp.Where) -> str: 1806 this = self.indent(self.sql(expression, "this")) 1807 return f"{self.seg('WHERE')}{self.sep()}{this}" 1808 1809 def window_sql(self, expression: exp.Window) -> str: 1810 this = self.sql(expression, "this") 1811 partition = self.partition_by_sql(expression) 1812 order = expression.args.get("order") 1813 order = self.order_sql(order, flat=True) if order else "" 1814 spec = self.sql(expression, "spec") 1815 alias = self.sql(expression, "alias") 1816 over = self.sql(expression, "over") or "OVER" 1817 1818 this = f"{this} {'AS' if expression.arg_key == 'windows' else over}" 1819 1820 first = expression.args.get("first") 1821 if first is None: 1822 first = "" 1823 else: 1824 first = "FIRST" if first else "LAST" 1825 1826 if not partition and not order and not spec and alias: 1827 return f"{this} {alias}" 1828 1829 args = " ".join(arg for arg in (alias, first, partition, order, spec) if arg) 1830 return f"{this} ({args})" 1831 1832 def partition_by_sql(self, expression: exp.Window | exp.MatchRecognize) -> str: 1833 partition = self.expressions(expression, key="partition_by", flat=True) 1834 return f"PARTITION BY {partition}" if partition else "" 1835 1836 def windowspec_sql(self, expression: exp.WindowSpec) -> str: 1837 kind = self.sql(expression, "kind") 1838 start = csv(self.sql(expression, "start"), self.sql(expression, "start_side"), sep=" ") 1839 end = ( 1840 csv(self.sql(expression, "end"), self.sql(expression, "end_side"), sep=" ") 1841 or "CURRENT ROW" 1842 ) 1843 return f"{kind} BETWEEN {start} AND {end}" 1844 1845 def withingroup_sql(self, expression: exp.WithinGroup) -> str: 1846 this = self.sql(expression, "this") 1847 expression_sql = self.sql(expression, "expression")[1:] # order has a leading space 1848 return f"{this} WITHIN GROUP ({expression_sql})" 1849 1850 def between_sql(self, expression: exp.Between) -> str: 1851 this = self.sql(expression, "this") 1852 low = self.sql(expression, "low") 1853 high = self.sql(expression, "high") 1854 return f"{this} BETWEEN {low} AND {high}" 1855 1856 def bracket_sql(self, expression: exp.Bracket) -> str: 1857 expressions = apply_index_offset(expression.this, expression.expressions, self.INDEX_OFFSET) 1858 expressions_sql = ", ".join(self.sql(e) for e in expressions) 1859 1860 return f"{self.sql(expression, 'this')}[{expressions_sql}]" 1861 1862 def safebracket_sql(self, expression: exp.SafeBracket) -> str: 1863 return self.bracket_sql(expression) 1864 1865 def all_sql(self, expression: exp.All) -> str: 1866 return f"ALL {self.wrap(expression)}" 1867 1868 def any_sql(self, expression: exp.Any) -> str: 1869 this = self.sql(expression, "this") 1870 if isinstance(expression.this, exp.Subqueryable): 1871 this = self.wrap(this) 1872 return f"ANY {this}" 1873 1874 def exists_sql(self, expression: exp.Exists) -> str: 1875 return f"EXISTS{self.wrap(expression)}" 1876 1877 def case_sql(self, expression: exp.Case) -> str: 1878 this = self.sql(expression, "this") 1879 statements = [f"CASE {this}" if this else "CASE"] 1880 1881 for e in expression.args["ifs"]: 1882 statements.append(f"WHEN {self.sql(e, 'this')}") 1883 statements.append(f"THEN {self.sql(e, 'true')}") 1884 1885 default = self.sql(expression, "default") 1886 1887 if default: 1888 statements.append(f"ELSE {default}") 1889 1890 statements.append("END") 1891 1892 if self.pretty and self.text_width(statements) > self.max_text_width: 1893 return self.indent("\n".join(statements), skip_first=True, skip_last=True) 1894 1895 return " ".join(statements) 1896 1897 def constraint_sql(self, expression: exp.Constraint) -> str: 1898 this = self.sql(expression, "this") 1899 expressions = self.expressions(expression, flat=True) 1900 return f"CONSTRAINT {this} {expressions}" 1901 1902 def nextvaluefor_sql(self, expression: exp.NextValueFor) -> str: 1903 order = expression.args.get("order") 1904 order = f" OVER ({self.order_sql(order, flat=True)})" if order else "" 1905 return f"NEXT VALUE FOR {self.sql(expression, 'this')}{order}" 1906 1907 def extract_sql(self, expression: exp.Extract) -> str: 1908 this = self.sql(expression, "this") 1909 expression_sql = self.sql(expression, "expression") 1910 return f"EXTRACT({this} FROM {expression_sql})" 1911 1912 def trim_sql(self, expression: exp.Trim) -> str: 1913 trim_type = self.sql(expression, "position") 1914 1915 if trim_type == "LEADING": 1916 return self.func("LTRIM", expression.this) 1917 elif trim_type == "TRAILING": 1918 return self.func("RTRIM", expression.this) 1919 else: 1920 return self.func("TRIM", expression.this, expression.expression) 1921 1922 def safeconcat_sql(self, expression: exp.SafeConcat) -> str: 1923 expressions = expression.expressions 1924 if self.STRICT_STRING_CONCAT: 1925 expressions = (exp.cast(e, "text") for e in expressions) 1926 return self.func("CONCAT", *expressions) 1927 1928 def check_sql(self, expression: exp.Check) -> str: 1929 this = self.sql(expression, key="this") 1930 return f"CHECK ({this})" 1931 1932 def foreignkey_sql(self, expression: exp.ForeignKey) -> str: 1933 expressions = self.expressions(expression, flat=True) 1934 reference = self.sql(expression, "reference") 1935 reference = f" {reference}" if reference else "" 1936 delete = self.sql(expression, "delete") 1937 delete = f" ON DELETE {delete}" if delete else "" 1938 update = self.sql(expression, "update") 1939 update = f" ON UPDATE {update}" if update else "" 1940 return f"FOREIGN KEY ({expressions}){reference}{delete}{update}" 1941 1942 def primarykey_sql(self, expression: exp.ForeignKey) -> str: 1943 expressions = self.expressions(expression, flat=True) 1944 options = self.expressions(expression, key="options", flat=True, sep=" ") 1945 options = f" {options}" if options else "" 1946 return f"PRIMARY KEY ({expressions}){options}" 1947 1948 def if_sql(self, expression: exp.If) -> str: 1949 return self.case_sql(exp.Case(ifs=[expression], default=expression.args.get("false"))) 1950 1951 def matchagainst_sql(self, expression: exp.MatchAgainst) -> str: 1952 modifier = expression.args.get("modifier") 1953 modifier = f" {modifier}" if modifier else "" 1954 return f"{self.func('MATCH', *expression.expressions)} AGAINST({self.sql(expression, 'this')}{modifier})" 1955 1956 def jsonkeyvalue_sql(self, expression: exp.JSONKeyValue) -> str: 1957 return f"{self.sql(expression, 'this')}: {self.sql(expression, 'expression')}" 1958 1959 def jsonobject_sql(self, expression: exp.JSONObject) -> str: 1960 null_handling = expression.args.get("null_handling") 1961 null_handling = f" {null_handling}" if null_handling else "" 1962 unique_keys = expression.args.get("unique_keys") 1963 if unique_keys is not None: 1964 unique_keys = f" {'WITH' if unique_keys else 'WITHOUT'} UNIQUE KEYS" 1965 else: 1966 unique_keys = "" 1967 return_type = self.sql(expression, "return_type") 1968 return_type = f" RETURNING {return_type}" if return_type else "" 1969 format_json = " FORMAT JSON" if expression.args.get("format_json") else "" 1970 encoding = self.sql(expression, "encoding") 1971 encoding = f" ENCODING {encoding}" if encoding else "" 1972 return self.func( 1973 "JSON_OBJECT", 1974 *expression.expressions, 1975 suffix=f"{null_handling}{unique_keys}{return_type}{format_json}{encoding})", 1976 ) 1977 1978 def openjsoncolumndef_sql(self, expression: exp.OpenJSONColumnDef) -> str: 1979 this = self.sql(expression, "this") 1980 kind = self.sql(expression, "kind") 1981 path = self.sql(expression, "path") 1982 path = f" {path}" if path else "" 1983 as_json = " AS JSON" if expression.args.get("as_json") else "" 1984 return f"{this} {kind}{path}{as_json}" 1985 1986 def openjson_sql(self, expression: exp.OpenJSON) -> str: 1987 this = self.sql(expression, "this") 1988 path = self.sql(expression, "path") 1989 path = f", {path}" if path else "" 1990 expressions = self.expressions(expression) 1991 with_ = ( 1992 f" WITH ({self.seg(self.indent(expressions), sep='')}{self.seg(')', sep='')}" 1993 if expressions 1994 else "" 1995 ) 1996 return f"OPENJSON({this}{path}){with_}" 1997 1998 def in_sql(self, expression: exp.In) -> str: 1999 query = expression.args.get("query") 2000 unnest = expression.args.get("unnest") 2001 field = expression.args.get("field") 2002 is_global = " GLOBAL" if expression.args.get("is_global") else "" 2003 2004 if query: 2005 in_sql = self.wrap(query) 2006 elif unnest: 2007 in_sql = self.in_unnest_op(unnest) 2008 elif field: 2009 in_sql = self.sql(field) 2010 else: 2011 in_sql = f"({self.expressions(expression, flat=True)})" 2012 2013 return f"{self.sql(expression, 'this')}{is_global} IN {in_sql}" 2014 2015 def in_unnest_op(self, unnest: exp.Unnest) -> str: 2016 return f"(SELECT {self.sql(unnest)})" 2017 2018 def interval_sql(self, expression: exp.Interval) -> str: 2019 unit = self.sql(expression, "unit") 2020 if not self.INTERVAL_ALLOWS_PLURAL_FORM: 2021 unit = self.TIME_PART_SINGULARS.get(unit.lower(), unit) 2022 unit = f" {unit}" if unit else "" 2023 2024 if self.SINGLE_STRING_INTERVAL: 2025 this = expression.this.name if expression.this else "" 2026 return f"INTERVAL '{this}{unit}'" if this else f"INTERVAL{unit}" 2027 2028 this = self.sql(expression, "this") 2029 if this: 2030 unwrapped = isinstance(expression.this, self.UNWRAPPED_INTERVAL_VALUES) 2031 this = f" {this}" if unwrapped else f" ({this})" 2032 2033 return f"INTERVAL{this}{unit}" 2034 2035 def return_sql(self, expression: exp.Return) -> str: 2036 return f"RETURN {self.sql(expression, 'this')}" 2037 2038 def reference_sql(self, expression: exp.Reference) -> str: 2039 this = self.sql(expression, "this") 2040 expressions = self.expressions(expression, flat=True) 2041 expressions = f"({expressions})" if expressions else "" 2042 options = self.expressions(expression, key="options", flat=True, sep=" ") 2043 options = f" {options}" if options else "" 2044 return f"REFERENCES {this}{expressions}{options}" 2045 2046 def anonymous_sql(self, expression: exp.Anonymous) -> str: 2047 return self.func(expression.name, *expression.expressions) 2048 2049 def paren_sql(self, expression: exp.Paren) -> str: 2050 if isinstance(expression.unnest(), exp.Select): 2051 sql = self.wrap(expression) 2052 else: 2053 sql = self.seg(self.indent(self.sql(expression, "this")), sep="") 2054 sql = f"({sql}{self.seg(')', sep='')}" 2055 2056 return self.prepend_ctes(expression, sql) 2057 2058 def neg_sql(self, expression: exp.Neg) -> str: 2059 # This makes sure we don't convert "- - 5" to "--5", which is a comment 2060 this_sql = self.sql(expression, "this") 2061 sep = " " if this_sql[0] == "-" else "" 2062 return f"-{sep}{this_sql}" 2063 2064 def not_sql(self, expression: exp.Not) -> str: 2065 return f"NOT {self.sql(expression, 'this')}" 2066 2067 def alias_sql(self, expression: exp.Alias) -> str: 2068 alias = self.sql(expression, "alias") 2069 alias = f" AS {alias}" if alias else "" 2070 return f"{self.sql(expression, 'this')}{alias}" 2071 2072 def aliases_sql(self, expression: exp.Aliases) -> str: 2073 return f"{self.sql(expression, 'this')} AS ({self.expressions(expression, flat=True)})" 2074 2075 def attimezone_sql(self, expression: exp.AtTimeZone) -> str: 2076 this = self.sql(expression, "this") 2077 zone = self.sql(expression, "zone") 2078 return f"{this} AT TIME ZONE {zone}" 2079 2080 def add_sql(self, expression: exp.Add) -> str: 2081 return self.binary(expression, "+") 2082 2083 def and_sql(self, expression: exp.And) -> str: 2084 return self.connector_sql(expression, "AND") 2085 2086 def xor_sql(self, expression: exp.Xor) -> str: 2087 return self.connector_sql(expression, "XOR") 2088 2089 def connector_sql(self, expression: exp.Connector, op: str) -> str: 2090 if not self.pretty: 2091 return self.binary(expression, op) 2092 2093 sqls = tuple( 2094 self.maybe_comment(self.sql(e), e, e.parent.comments or []) if i != 1 else self.sql(e) 2095 for i, e in enumerate(expression.flatten(unnest=False)) 2096 ) 2097 2098 sep = "\n" if self.text_width(sqls) > self.max_text_width else " " 2099 return f"{sep}{op} ".join(sqls) 2100 2101 def bitwiseand_sql(self, expression: exp.BitwiseAnd) -> str: 2102 return self.binary(expression, "&") 2103 2104 def bitwiseleftshift_sql(self, expression: exp.BitwiseLeftShift) -> str: 2105 return self.binary(expression, "<<") 2106 2107 def bitwisenot_sql(self, expression: exp.BitwiseNot) -> str: 2108 return f"~{self.sql(expression, 'this')}" 2109 2110 def bitwiseor_sql(self, expression: exp.BitwiseOr) -> str: 2111 return self.binary(expression, "|") 2112 2113 def bitwiserightshift_sql(self, expression: exp.BitwiseRightShift) -> str: 2114 return self.binary(expression, ">>") 2115 2116 def bitwisexor_sql(self, expression: exp.BitwiseXor) -> str: 2117 return self.binary(expression, "^") 2118 2119 def cast_sql(self, expression: exp.Cast, safe_prefix: t.Optional[str] = None) -> str: 2120 format_sql = self.sql(expression, "format") 2121 format_sql = f" FORMAT {format_sql}" if format_sql else "" 2122 return f"{safe_prefix or ''}CAST({self.sql(expression, 'this')} AS {self.sql(expression, 'to')}{format_sql})" 2123 2124 def currentdate_sql(self, expression: exp.CurrentDate) -> str: 2125 zone = self.sql(expression, "this") 2126 return f"CURRENT_DATE({zone})" if zone else "CURRENT_DATE" 2127 2128 def collate_sql(self, expression: exp.Collate) -> str: 2129 return self.binary(expression, "COLLATE") 2130 2131 def command_sql(self, expression: exp.Command) -> str: 2132 return f"{self.sql(expression, 'this').upper()} {expression.text('expression').strip()}" 2133 2134 def comment_sql(self, expression: exp.Comment) -> str: 2135 this = self.sql(expression, "this") 2136 kind = expression.args["kind"] 2137 exists_sql = " IF EXISTS " if expression.args.get("exists") else " " 2138 expression_sql = self.sql(expression, "expression") 2139 return f"COMMENT{exists_sql}ON {kind} {this} IS {expression_sql}" 2140 2141 def mergetreettlaction_sql(self, expression: exp.MergeTreeTTLAction) -> str: 2142 this = self.sql(expression, "this") 2143 delete = " DELETE" if expression.args.get("delete") else "" 2144 recompress = self.sql(expression, "recompress") 2145 recompress = f" RECOMPRESS {recompress}" if recompress else "" 2146 to_disk = self.sql(expression, "to_disk") 2147 to_disk = f" TO DISK {to_disk}" if to_disk else "" 2148 to_volume = self.sql(expression, "to_volume") 2149 to_volume = f" TO VOLUME {to_volume}" if to_volume else "" 2150 return f"{this}{delete}{recompress}{to_disk}{to_volume}" 2151 2152 def mergetreettl_sql(self, expression: exp.MergeTreeTTL) -> str: 2153 where = self.sql(expression, "where") 2154 group = self.sql(expression, "group") 2155 aggregates = self.expressions(expression, key="aggregates") 2156 aggregates = self.seg("SET") + self.seg(aggregates) if aggregates else "" 2157 2158 if not (where or group or aggregates) and len(expression.expressions) == 1: 2159 return f"TTL {self.expressions(expression, flat=True)}" 2160 2161 return f"TTL{self.seg(self.expressions(expression))}{where}{group}{aggregates}" 2162 2163 def transaction_sql(self, expression: exp.Transaction) -> str: 2164 return "BEGIN" 2165 2166 def commit_sql(self, expression: exp.Commit) -> str: 2167 chain = expression.args.get("chain") 2168 if chain is not None: 2169 chain = " AND CHAIN" if chain else " AND NO CHAIN" 2170 2171 return f"COMMIT{chain or ''}" 2172 2173 def rollback_sql(self, expression: exp.Rollback) -> str: 2174 savepoint = expression.args.get("savepoint") 2175 savepoint = f" TO {savepoint}" if savepoint else "" 2176 return f"ROLLBACK{savepoint}" 2177 2178 def altercolumn_sql(self, expression: exp.AlterColumn) -> str: 2179 this = self.sql(expression, "this") 2180 2181 dtype = self.sql(expression, "dtype") 2182 if dtype: 2183 collate = self.sql(expression, "collate") 2184 collate = f" COLLATE {collate}" if collate else "" 2185 using = self.sql(expression, "using") 2186 using = f" USING {using}" if using else "" 2187 return f"ALTER COLUMN {this} TYPE {dtype}{collate}{using}" 2188 2189 default = self.sql(expression, "default") 2190 if default: 2191 return f"ALTER COLUMN {this} SET DEFAULT {default}" 2192 2193 if not expression.args.get("drop"): 2194 self.unsupported("Unsupported ALTER COLUMN syntax") 2195 2196 return f"ALTER COLUMN {this} DROP DEFAULT" 2197 2198 def renametable_sql(self, expression: exp.RenameTable) -> str: 2199 if not self.RENAME_TABLE_WITH_DB: 2200 # Remove db from tables 2201 expression = expression.transform( 2202 lambda n: exp.table_(n.this) if isinstance(n, exp.Table) else n 2203 ) 2204 this = self.sql(expression, "this") 2205 return f"RENAME TO {this}" 2206 2207 def altertable_sql(self, expression: exp.AlterTable) -> str: 2208 actions = expression.args["actions"] 2209 2210 if isinstance(actions[0], exp.ColumnDef): 2211 actions = self.expressions(expression, key="actions", prefix="ADD COLUMN ") 2212 elif isinstance(actions[0], exp.Schema): 2213 actions = self.expressions(expression, key="actions", prefix="ADD COLUMNS ") 2214 elif isinstance(actions[0], exp.Delete): 2215 actions = self.expressions(expression, key="actions", flat=True) 2216 else: 2217 actions = self.expressions(expression, key="actions") 2218 2219 exists = " IF EXISTS" if expression.args.get("exists") else "" 2220 return f"ALTER TABLE{exists} {self.sql(expression, 'this')} {actions}" 2221 2222 def droppartition_sql(self, expression: exp.DropPartition) -> str: 2223 expressions = self.expressions(expression) 2224 exists = " IF EXISTS " if expression.args.get("exists") else " " 2225 return f"DROP{exists}{expressions}" 2226 2227 def addconstraint_sql(self, expression: exp.AddConstraint) -> str: 2228 this = self.sql(expression, "this") 2229 expression_ = self.sql(expression, "expression") 2230 add_constraint = f"ADD CONSTRAINT {this}" if this else "ADD" 2231 2232 enforced = expression.args.get("enforced") 2233 if enforced is not None: 2234 return f"{add_constraint} CHECK ({expression_}){' ENFORCED' if enforced else ''}" 2235 2236 return f"{add_constraint} {expression_}" 2237 2238 def distinct_sql(self, expression: exp.Distinct) -> str: 2239 this = self.expressions(expression, flat=True) 2240 this = f" {this}" if this else "" 2241 2242 on = self.sql(expression, "on") 2243 on = f" ON {on}" if on else "" 2244 return f"DISTINCT{this}{on}" 2245 2246 def ignorenulls_sql(self, expression: exp.IgnoreNulls) -> str: 2247 return f"{self.sql(expression, 'this')} IGNORE NULLS" 2248 2249 def respectnulls_sql(self, expression: exp.RespectNulls) -> str: 2250 return f"{self.sql(expression, 'this')} RESPECT NULLS" 2251 2252 def intdiv_sql(self, expression: exp.IntDiv) -> str: 2253 return self.sql( 2254 exp.Cast( 2255 this=exp.Div(this=expression.this, expression=expression.expression), 2256 to=exp.DataType(this=exp.DataType.Type.INT), 2257 ) 2258 ) 2259 2260 def dpipe_sql(self, expression: exp.DPipe) -> str: 2261 return self.binary(expression, "||") 2262 2263 def safedpipe_sql(self, expression: exp.SafeDPipe) -> str: 2264 if self.STRICT_STRING_CONCAT: 2265 return self.func("CONCAT", *(exp.cast(e, "text") for e in expression.flatten())) 2266 return self.dpipe_sql(expression) 2267 2268 def div_sql(self, expression: exp.Div) -> str: 2269 return self.binary(expression, "/") 2270 2271 def overlaps_sql(self, expression: exp.Overlaps) -> str: 2272 return self.binary(expression, "OVERLAPS") 2273 2274 def distance_sql(self, expression: exp.Distance) -> str: 2275 return self.binary(expression, "<->") 2276 2277 def dot_sql(self, expression: exp.Dot) -> str: 2278 return f"{self.sql(expression, 'this')}.{self.sql(expression, 'expression')}" 2279 2280 def eq_sql(self, expression: exp.EQ) -> str: 2281 return self.binary(expression, "=") 2282 2283 def escape_sql(self, expression: exp.Escape) -> str: 2284 return self.binary(expression, "ESCAPE") 2285 2286 def glob_sql(self, expression: exp.Glob) -> str: 2287 return self.binary(expression, "GLOB") 2288 2289 def gt_sql(self, expression: exp.GT) -> str: 2290 return self.binary(expression, ">") 2291 2292 def gte_sql(self, expression: exp.GTE) -> str: 2293 return self.binary(expression, ">=") 2294 2295 def ilike_sql(self, expression: exp.ILike) -> str: 2296 return self.binary(expression, "ILIKE") 2297 2298 def ilikeany_sql(self, expression: exp.ILikeAny) -> str: 2299 return self.binary(expression, "ILIKE ANY") 2300 2301 def is_sql(self, expression: exp.Is) -> str: 2302 if not self.IS_BOOL_ALLOWED and isinstance(expression.expression, exp.Boolean): 2303 return self.sql( 2304 expression.this if expression.expression.this else exp.not_(expression.this) 2305 ) 2306 return self.binary(expression, "IS") 2307 2308 def like_sql(self, expression: exp.Like) -> str: 2309 return self.binary(expression, "LIKE") 2310 2311 def likeany_sql(self, expression: exp.LikeAny) -> str: 2312 return self.binary(expression, "LIKE ANY") 2313 2314 def similarto_sql(self, expression: exp.SimilarTo) -> str: 2315 return self.binary(expression, "SIMILAR TO") 2316 2317 def lt_sql(self, expression: exp.LT) -> str: 2318 return self.binary(expression, "<") 2319 2320 def lte_sql(self, expression: exp.LTE) -> str: 2321 return self.binary(expression, "<=") 2322 2323 def mod_sql(self, expression: exp.Mod) -> str: 2324 return self.binary(expression, "%") 2325 2326 def mul_sql(self, expression: exp.Mul) -> str: 2327 return self.binary(expression, "*") 2328 2329 def neq_sql(self, expression: exp.NEQ) -> str: 2330 return self.binary(expression, "<>") 2331 2332 def nullsafeeq_sql(self, expression: exp.NullSafeEQ) -> str: 2333 return self.binary(expression, "IS NOT DISTINCT FROM") 2334 2335 def nullsafeneq_sql(self, expression: exp.NullSafeNEQ) -> str: 2336 return self.binary(expression, "IS DISTINCT FROM") 2337 2338 def or_sql(self, expression: exp.Or) -> str: 2339 return self.connector_sql(expression, "OR") 2340 2341 def slice_sql(self, expression: exp.Slice) -> str: 2342 return self.binary(expression, ":") 2343 2344 def sub_sql(self, expression: exp.Sub) -> str: 2345 return self.binary(expression, "-") 2346 2347 def trycast_sql(self, expression: exp.TryCast) -> str: 2348 return self.cast_sql(expression, safe_prefix="TRY_") 2349 2350 def use_sql(self, expression: exp.Use) -> str: 2351 kind = self.sql(expression, "kind") 2352 kind = f" {kind}" if kind else "" 2353 this = self.sql(expression, "this") 2354 this = f" {this}" if this else "" 2355 return f"USE{kind}{this}" 2356 2357 def binary(self, expression: exp.Binary, op: str) -> str: 2358 op = self.maybe_comment(op, comments=expression.comments) 2359 return f"{self.sql(expression, 'this')} {op} {self.sql(expression, 'expression')}" 2360 2361 def function_fallback_sql(self, expression: exp.Func) -> str: 2362 args = [] 2363 2364 for key in expression.arg_types: 2365 arg_value = expression.args.get(key) 2366 2367 if isinstance(arg_value, list): 2368 for value in arg_value: 2369 args.append(value) 2370 elif arg_value is not None: 2371 args.append(arg_value) 2372 2373 return self.func(expression.sql_name(), *args) 2374 2375 def func( 2376 self, 2377 name: str, 2378 *args: t.Optional[exp.Expression | str], 2379 prefix: str = "(", 2380 suffix: str = ")", 2381 ) -> str: 2382 return f"{self.normalize_func(name)}{prefix}{self.format_args(*args)}{suffix}" 2383 2384 def format_args(self, *args: t.Optional[str | exp.Expression]) -> str: 2385 arg_sqls = tuple(self.sql(arg) for arg in args if arg is not None) 2386 if self.pretty and self.text_width(arg_sqls) > self.max_text_width: 2387 return self.indent("\n" + f",\n".join(arg_sqls) + "\n", skip_first=True, skip_last=True) 2388 return ", ".join(arg_sqls) 2389 2390 def text_width(self, args: t.Iterable) -> int: 2391 return sum(len(arg) for arg in args) 2392 2393 def format_time(self, expression: exp.Expression) -> t.Optional[str]: 2394 return format_time( 2395 self.sql(expression, "format"), self.INVERSE_TIME_MAPPING, self.INVERSE_TIME_TRIE 2396 ) 2397 2398 def expressions( 2399 self, 2400 expression: t.Optional[exp.Expression] = None, 2401 key: t.Optional[str] = None, 2402 sqls: t.Optional[t.List[str]] = None, 2403 flat: bool = False, 2404 indent: bool = True, 2405 skip_first: bool = False, 2406 sep: str = ", ", 2407 prefix: str = "", 2408 ) -> str: 2409 expressions = expression.args.get(key or "expressions") if expression else sqls 2410 2411 if not expressions: 2412 return "" 2413 2414 if flat: 2415 return sep.join(self.sql(e) for e in expressions) 2416 2417 num_sqls = len(expressions) 2418 2419 # These are calculated once in case we have the leading_comma / pretty option set, correspondingly 2420 pad = " " * self.pad 2421 stripped_sep = sep.strip() 2422 2423 result_sqls = [] 2424 for i, e in enumerate(expressions): 2425 sql = self.sql(e, comment=False) 2426 comments = self.maybe_comment("", e) if isinstance(e, exp.Expression) else "" 2427 2428 if self.pretty: 2429 if self.leading_comma: 2430 result_sqls.append(f"{sep if i > 0 else pad}{prefix}{sql}{comments}") 2431 else: 2432 result_sqls.append( 2433 f"{prefix}{sql}{stripped_sep if i + 1 < num_sqls else ''}{comments}" 2434 ) 2435 else: 2436 result_sqls.append(f"{prefix}{sql}{comments}{sep if i + 1 < num_sqls else ''}") 2437 2438 result_sql = "\n".join(result_sqls) if self.pretty else "".join(result_sqls) 2439 return self.indent(result_sql, skip_first=skip_first) if indent else result_sql 2440 2441 def op_expressions(self, op: str, expression: exp.Expression, flat: bool = False) -> str: 2442 flat = flat or isinstance(expression.parent, exp.Properties) 2443 expressions_sql = self.expressions(expression, flat=flat) 2444 if flat: 2445 return f"{op} {expressions_sql}" 2446 return f"{self.seg(op)}{self.sep() if expressions_sql else ''}{expressions_sql}" 2447 2448 def naked_property(self, expression: exp.Property) -> str: 2449 property_name = exp.Properties.PROPERTY_TO_NAME.get(expression.__class__) 2450 if not property_name: 2451 self.unsupported(f"Unsupported property {expression.__class__.__name__}") 2452 return f"{property_name} {self.sql(expression, 'this')}" 2453 2454 def set_operation(self, expression: exp.Expression, op: str) -> str: 2455 this = self.sql(expression, "this") 2456 op = self.seg(op) 2457 return self.query_modifiers( 2458 expression, f"{this}{op}{self.sep()}{self.sql(expression, 'expression')}" 2459 ) 2460 2461 def tag_sql(self, expression: exp.Tag) -> str: 2462 return f"{expression.args.get('prefix')}{self.sql(expression.this)}{expression.args.get('postfix')}" 2463 2464 def token_sql(self, token_type: TokenType) -> str: 2465 return self.TOKEN_MAPPING.get(token_type, token_type.name) 2466 2467 def userdefinedfunction_sql(self, expression: exp.UserDefinedFunction) -> str: 2468 this = self.sql(expression, "this") 2469 expressions = self.no_identify(self.expressions, expression) 2470 expressions = ( 2471 self.wrap(expressions) if expression.args.get("wrapped") else f" {expressions}" 2472 ) 2473 return f"{this}{expressions}" 2474 2475 def joinhint_sql(self, expression: exp.JoinHint) -> str: 2476 this = self.sql(expression, "this") 2477 expressions = self.expressions(expression, flat=True) 2478 return f"{this}({expressions})" 2479 2480 def kwarg_sql(self, expression: exp.Kwarg) -> str: 2481 return self.binary(expression, "=>") 2482 2483 def when_sql(self, expression: exp.When) -> str: 2484 matched = "MATCHED" if expression.args["matched"] else "NOT MATCHED" 2485 source = " BY SOURCE" if self.MATCHED_BY_SOURCE and expression.args.get("source") else "" 2486 condition = self.sql(expression, "condition") 2487 condition = f" AND {condition}" if condition else "" 2488 2489 then_expression = expression.args.get("then") 2490 if isinstance(then_expression, exp.Insert): 2491 then = f"INSERT {self.sql(then_expression, 'this')}" 2492 if "expression" in then_expression.args: 2493 then += f" VALUES {self.sql(then_expression, 'expression')}" 2494 elif isinstance(then_expression, exp.Update): 2495 if isinstance(then_expression.args.get("expressions"), exp.Star): 2496 then = f"UPDATE {self.sql(then_expression, 'expressions')}" 2497 else: 2498 then = f"UPDATE SET {self.expressions(then_expression, flat=True)}" 2499 else: 2500 then = self.sql(then_expression) 2501 return f"WHEN {matched}{source}{condition} THEN {then}" 2502 2503 def merge_sql(self, expression: exp.Merge) -> str: 2504 this = self.sql(expression, "this") 2505 using = f"USING {self.sql(expression, 'using')}" 2506 on = f"ON {self.sql(expression, 'on')}" 2507 return f"MERGE INTO {this} {using} {on} {self.expressions(expression, sep=' ')}" 2508 2509 def tochar_sql(self, expression: exp.ToChar) -> str: 2510 if expression.args.get("format"): 2511 self.unsupported("Format argument unsupported for TO_CHAR/TO_VARCHAR function") 2512 2513 return self.sql(exp.cast(expression.this, "text")) 2514 2515 def dictproperty_sql(self, expression: exp.DictProperty) -> str: 2516 this = self.sql(expression, "this") 2517 kind = self.sql(expression, "kind") 2518 settings_sql = self.expressions(expression, key="settings", sep=" ") 2519 args = f"({self.sep('')}{settings_sql}{self.seg(')', sep='')}" if settings_sql else "()" 2520 return f"{this}({kind}{args})" 2521 2522 def dictrange_sql(self, expression: exp.DictRange) -> str: 2523 this = self.sql(expression, "this") 2524 max = self.sql(expression, "max") 2525 min = self.sql(expression, "min") 2526 return f"{this}(MIN {min} MAX {max})" 2527 2528 def dictsubproperty_sql(self, expression: exp.DictSubProperty) -> str: 2529 return f"{self.sql(expression, 'this')} {self.sql(expression, 'value')}" 2530 2531 def oncluster_sql(self, expression: exp.OnCluster) -> str: 2532 return "" 2533 2534 def clusteredbyproperty_sql(self, expression: exp.ClusteredByProperty) -> str: 2535 expressions = self.expressions(expression, key="expressions", flat=True) 2536 sorted_by = self.expressions(expression, key="sorted_by", flat=True) 2537 sorted_by = f" SORTED BY ({sorted_by})" if sorted_by else "" 2538 buckets = self.sql(expression, "buckets") 2539 return f"CLUSTERED BY ({expressions}){sorted_by} INTO {buckets} BUCKETS" 2540 2541 def anyvalue_sql(self, expression: exp.AnyValue) -> str: 2542 this = self.sql(expression, "this") 2543 having = self.sql(expression, "having") 2544 2545 if having: 2546 this = f"{this} HAVING {'MAX' if expression.args.get('max') else 'MIN'} {having}" 2547 2548 return self.func("ANY_VALUE", this) 2549 2550 def querytransform_sql(self, expression: exp.QueryTransform) -> str: 2551 transform = self.func("TRANSFORM", *expression.expressions) 2552 row_format_before = self.sql(expression, "row_format_before") 2553 row_format_before = f" {row_format_before}" if row_format_before else "" 2554 record_writer = self.sql(expression, "record_writer") 2555 record_writer = f" RECORDWRITER {record_writer}" if record_writer else "" 2556 using = f" USING {self.sql(expression, 'command_script')}" 2557 schema = self.sql(expression, "schema") 2558 schema = f" AS {schema}" if schema else "" 2559 row_format_after = self.sql(expression, "row_format_after") 2560 row_format_after = f" {row_format_after}" if row_format_after else "" 2561 record_reader = self.sql(expression, "record_reader") 2562 record_reader = f" RECORDREADER {record_reader}" if record_reader else "" 2563 return f"{transform}{row_format_before}{record_writer}{using}{schema}{row_format_after}{record_reader}" 2564 2565 2566def cached_generator( 2567 cache: t.Optional[t.Dict[int, str]] = None 2568) -> t.Callable[[exp.Expression], str]: 2569 """Returns a cached generator.""" 2570 cache = {} if cache is None else cache 2571 generator = Generator(normalize=True, identify="safe") 2572 return lambda e: generator.generate(e, cache)
logger =
<Logger sqlglot (WARNING)>
class
Generator:
16class Generator: 17 """ 18 Generator converts a given syntax tree to the corresponding SQL string. 19 20 Args: 21 pretty: Whether or not to format the produced SQL string. 22 Default: False. 23 identify: Determines when an identifier should be quoted. Possible values are: 24 False (default): Never quote, except in cases where it's mandatory by the dialect. 25 True or 'always': Always quote. 26 'safe': Only quote identifiers that are case insensitive. 27 normalize: Whether or not to normalize identifiers to lowercase. 28 Default: False. 29 pad: Determines the pad size in a formatted string. 30 Default: 2. 31 indent: Determines the indentation size in a formatted string. 32 Default: 2. 33 normalize_functions: Whether or not to normalize all function names. Possible values are: 34 "upper" or True (default): Convert names to uppercase. 35 "lower": Convert names to lowercase. 36 False: Disables function name normalization. 37 unsupported_level: Determines the generator's behavior when it encounters unsupported expressions. 38 Default ErrorLevel.WARN. 39 max_unsupported: Maximum number of unsupported messages to include in a raised UnsupportedError. 40 This is only relevant if unsupported_level is ErrorLevel.RAISE. 41 Default: 3 42 leading_comma: Determines whether or not the comma is leading or trailing in select expressions. 43 This is only relevant when generating in pretty mode. 44 Default: False 45 max_text_width: The max number of characters in a segment before creating new lines in pretty mode. 46 The default is on the smaller end because the length only represents a segment and not the true 47 line length. 48 Default: 80 49 comments: Whether or not to preserve comments in the output SQL code. 50 Default: True 51 """ 52 53 TRANSFORMS = { 54 exp.DateAdd: lambda self, e: self.func( 55 "DATE_ADD", e.this, e.expression, exp.Literal.string(e.text("unit")) 56 ), 57 exp.TsOrDsAdd: lambda self, e: self.func( 58 "TS_OR_DS_ADD", e.this, e.expression, exp.Literal.string(e.text("unit")) 59 ), 60 exp.CaseSpecificColumnConstraint: lambda self, e: f"{'NOT ' if e.args.get('not_') else ''}CASESPECIFIC", 61 exp.CharacterSetColumnConstraint: lambda self, e: f"CHARACTER SET {self.sql(e, 'this')}", 62 exp.CharacterSetProperty: lambda self, e: f"{'DEFAULT ' if e.args.get('default') else ''}CHARACTER SET={self.sql(e, 'this')}", 63 exp.CheckColumnConstraint: lambda self, e: f"CHECK ({self.sql(e, 'this')})", 64 exp.CollateColumnConstraint: lambda self, e: f"COLLATE {self.sql(e, 'this')}", 65 exp.CopyGrantsProperty: lambda self, e: "COPY GRANTS", 66 exp.CommentColumnConstraint: lambda self, e: f"COMMENT {self.sql(e, 'this')}", 67 exp.DateFormatColumnConstraint: lambda self, e: f"FORMAT {self.sql(e, 'this')}", 68 exp.DefaultColumnConstraint: lambda self, e: f"DEFAULT {self.sql(e, 'this')}", 69 exp.EncodeColumnConstraint: lambda self, e: f"ENCODE {self.sql(e, 'this')}", 70 exp.ExecuteAsProperty: lambda self, e: self.naked_property(e), 71 exp.ExternalProperty: lambda self, e: "EXTERNAL", 72 exp.InlineLengthColumnConstraint: lambda self, e: f"INLINE LENGTH {self.sql(e, 'this')}", 73 exp.LanguageProperty: lambda self, e: self.naked_property(e), 74 exp.LocationProperty: lambda self, e: self.naked_property(e), 75 exp.LogProperty: lambda self, e: f"{'NO ' if e.args.get('no') else ''}LOG", 76 exp.MaterializedProperty: lambda self, e: "MATERIALIZED", 77 exp.NoPrimaryIndexProperty: lambda self, e: "NO PRIMARY INDEX", 78 exp.OnCommitProperty: lambda self, e: f"ON COMMIT {'DELETE' if e.args.get('delete') else 'PRESERVE'} ROWS", 79 exp.OnUpdateColumnConstraint: lambda self, e: f"ON UPDATE {self.sql(e, 'this')}", 80 exp.PathColumnConstraint: lambda self, e: f"PATH {self.sql(e, 'this')}", 81 exp.ReturnsProperty: lambda self, e: self.naked_property(e), 82 exp.SetProperty: lambda self, e: f"{'MULTI' if e.args.get('multi') else ''}SET", 83 exp.SettingsProperty: lambda self, e: f"SETTINGS{self.seg('')}{(self.expressions(e))}", 84 exp.SqlSecurityProperty: lambda self, e: f"SQL SECURITY {'DEFINER' if e.args.get('definer') else 'INVOKER'}", 85 exp.StabilityProperty: lambda self, e: e.name, 86 exp.TemporaryProperty: lambda self, e: f"TEMPORARY", 87 exp.ToTableProperty: lambda self, e: f"TO {self.sql(e.this)}", 88 exp.TransientProperty: lambda self, e: "TRANSIENT", 89 exp.TitleColumnConstraint: lambda self, e: f"TITLE {self.sql(e, 'this')}", 90 exp.UppercaseColumnConstraint: lambda self, e: f"UPPERCASE", 91 exp.VarMap: lambda self, e: self.func("MAP", e.args["keys"], e.args["values"]), 92 exp.VolatileProperty: lambda self, e: "VOLATILE", 93 exp.WithJournalTableProperty: lambda self, e: f"WITH JOURNAL TABLE={self.sql(e, 'this')}", 94 } 95 96 # Whether or not null ordering is supported in order by 97 NULL_ORDERING_SUPPORTED = True 98 99 # Whether or not locking reads (i.e. SELECT ... FOR UPDATE/SHARE) are supported 100 LOCKING_READS_SUPPORTED = False 101 102 # Always do union distinct or union all 103 EXPLICIT_UNION = False 104 105 # Wrap derived values in parens, usually standard but spark doesn't support it 106 WRAP_DERIVED_VALUES = True 107 108 # Whether or not create function uses an AS before the RETURN 109 CREATE_FUNCTION_RETURN_AS = True 110 111 # Whether or not MERGE ... WHEN MATCHED BY SOURCE is allowed 112 MATCHED_BY_SOURCE = True 113 114 # Whether or not the INTERVAL expression works only with values like '1 day' 115 SINGLE_STRING_INTERVAL = False 116 117 # Whether or not the plural form of date parts like day (i.e. "days") is supported in INTERVALs 118 INTERVAL_ALLOWS_PLURAL_FORM = True 119 120 # Whether or not the TABLESAMPLE clause supports a method name, like BERNOULLI 121 TABLESAMPLE_WITH_METHOD = True 122 123 # Whether or not to treat the number in TABLESAMPLE (50) as a percentage 124 TABLESAMPLE_SIZE_IS_PERCENT = False 125 126 # Whether or not limit and fetch are supported (possible values: "ALL", "LIMIT", "FETCH") 127 LIMIT_FETCH = "ALL" 128 129 # Whether or not a table is allowed to be renamed with a db 130 RENAME_TABLE_WITH_DB = True 131 132 # The separator for grouping sets and rollups 133 GROUPINGS_SEP = "," 134 135 # The string used for creating an index on a table 136 INDEX_ON = "ON" 137 138 # Whether or not join hints should be generated 139 JOIN_HINTS = True 140 141 # Whether or not table hints should be generated 142 TABLE_HINTS = True 143 144 # Whether or not query hints should be generated 145 QUERY_HINTS = True 146 147 # What kind of separator to use for query hints 148 QUERY_HINT_SEP = ", " 149 150 # Whether or not comparing against booleans (e.g. x IS TRUE) is supported 151 IS_BOOL_ALLOWED = True 152 153 # Whether or not to include the "SET" keyword in the "INSERT ... ON DUPLICATE KEY UPDATE" statement 154 DUPLICATE_KEY_UPDATE_WITH_SET = True 155 156 # Whether or not to generate the limit as TOP <value> instead of LIMIT <value> 157 LIMIT_IS_TOP = False 158 159 # Whether or not to generate INSERT INTO ... RETURNING or INSERT INTO RETURNING ... 160 RETURNING_END = True 161 162 # Whether or not to generate the (+) suffix for columns used in old-style join conditions 163 COLUMN_JOIN_MARKS_SUPPORTED = False 164 165 # https://cloud.google.com/bigquery/docs/reference/standard-sql/query-syntax 166 SELECT_KINDS: t.Tuple[str, ...] = ("STRUCT", "VALUE") 167 168 # Whether or not VALUES statements can be used as derived tables. 169 # MySQL 5 and Redshift do not allow this, so when False, it will convert 170 # SELECT * VALUES into SELECT UNION 171 VALUES_AS_TABLE = True 172 173 TYPE_MAPPING = { 174 exp.DataType.Type.NCHAR: "CHAR", 175 exp.DataType.Type.NVARCHAR: "VARCHAR", 176 exp.DataType.Type.MEDIUMTEXT: "TEXT", 177 exp.DataType.Type.LONGTEXT: "TEXT", 178 exp.DataType.Type.MEDIUMBLOB: "BLOB", 179 exp.DataType.Type.LONGBLOB: "BLOB", 180 exp.DataType.Type.INET: "INET", 181 } 182 183 STAR_MAPPING = { 184 "except": "EXCEPT", 185 "replace": "REPLACE", 186 } 187 188 TIME_PART_SINGULARS = { 189 "microseconds": "microsecond", 190 "seconds": "second", 191 "minutes": "minute", 192 "hours": "hour", 193 "days": "day", 194 "weeks": "week", 195 "months": "month", 196 "quarters": "quarter", 197 "years": "year", 198 } 199 200 TOKEN_MAPPING: t.Dict[TokenType, str] = {} 201 202 STRUCT_DELIMITER = ("<", ">") 203 204 PARAMETER_TOKEN = "@" 205 206 PROPERTIES_LOCATION = { 207 exp.AlgorithmProperty: exp.Properties.Location.POST_CREATE, 208 exp.AutoIncrementProperty: exp.Properties.Location.POST_SCHEMA, 209 exp.BlockCompressionProperty: exp.Properties.Location.POST_NAME, 210 exp.CharacterSetProperty: exp.Properties.Location.POST_SCHEMA, 211 exp.ChecksumProperty: exp.Properties.Location.POST_NAME, 212 exp.CollateProperty: exp.Properties.Location.POST_SCHEMA, 213 exp.CopyGrantsProperty: exp.Properties.Location.POST_SCHEMA, 214 exp.Cluster: exp.Properties.Location.POST_SCHEMA, 215 exp.ClusteredByProperty: exp.Properties.Location.POST_SCHEMA, 216 exp.DataBlocksizeProperty: exp.Properties.Location.POST_NAME, 217 exp.DefinerProperty: exp.Properties.Location.POST_CREATE, 218 exp.DictRange: exp.Properties.Location.POST_SCHEMA, 219 exp.DictProperty: exp.Properties.Location.POST_SCHEMA, 220 exp.DistKeyProperty: exp.Properties.Location.POST_SCHEMA, 221 exp.DistStyleProperty: exp.Properties.Location.POST_SCHEMA, 222 exp.EngineProperty: exp.Properties.Location.POST_SCHEMA, 223 exp.ExecuteAsProperty: exp.Properties.Location.POST_SCHEMA, 224 exp.ExternalProperty: exp.Properties.Location.POST_CREATE, 225 exp.FallbackProperty: exp.Properties.Location.POST_NAME, 226 exp.FileFormatProperty: exp.Properties.Location.POST_WITH, 227 exp.FreespaceProperty: exp.Properties.Location.POST_NAME, 228 exp.IsolatedLoadingProperty: exp.Properties.Location.POST_NAME, 229 exp.JournalProperty: exp.Properties.Location.POST_NAME, 230 exp.LanguageProperty: exp.Properties.Location.POST_SCHEMA, 231 exp.LikeProperty: exp.Properties.Location.POST_SCHEMA, 232 exp.LocationProperty: exp.Properties.Location.POST_SCHEMA, 233 exp.LockingProperty: exp.Properties.Location.POST_ALIAS, 234 exp.LogProperty: exp.Properties.Location.POST_NAME, 235 exp.MaterializedProperty: exp.Properties.Location.POST_CREATE, 236 exp.MergeBlockRatioProperty: exp.Properties.Location.POST_NAME, 237 exp.NoPrimaryIndexProperty: exp.Properties.Location.POST_EXPRESSION, 238 exp.OnCommitProperty: exp.Properties.Location.POST_EXPRESSION, 239 exp.Order: exp.Properties.Location.POST_SCHEMA, 240 exp.PartitionedByProperty: exp.Properties.Location.POST_WITH, 241 exp.PrimaryKey: exp.Properties.Location.POST_SCHEMA, 242 exp.Property: exp.Properties.Location.POST_WITH, 243 exp.ReturnsProperty: exp.Properties.Location.POST_SCHEMA, 244 exp.RowFormatProperty: exp.Properties.Location.POST_SCHEMA, 245 exp.RowFormatDelimitedProperty: exp.Properties.Location.POST_SCHEMA, 246 exp.RowFormatSerdeProperty: exp.Properties.Location.POST_SCHEMA, 247 exp.SchemaCommentProperty: exp.Properties.Location.POST_SCHEMA, 248 exp.SerdeProperties: exp.Properties.Location.POST_SCHEMA, 249 exp.Set: exp.Properties.Location.POST_SCHEMA, 250 exp.SettingsProperty: exp.Properties.Location.POST_SCHEMA, 251 exp.SetProperty: exp.Properties.Location.POST_CREATE, 252 exp.SortKeyProperty: exp.Properties.Location.POST_SCHEMA, 253 exp.SqlSecurityProperty: exp.Properties.Location.POST_CREATE, 254 exp.StabilityProperty: exp.Properties.Location.POST_SCHEMA, 255 exp.TemporaryProperty: exp.Properties.Location.POST_CREATE, 256 exp.ToTableProperty: exp.Properties.Location.POST_SCHEMA, 257 exp.TransientProperty: exp.Properties.Location.POST_CREATE, 258 exp.MergeTreeTTL: exp.Properties.Location.POST_SCHEMA, 259 exp.VolatileProperty: exp.Properties.Location.POST_CREATE, 260 exp.WithDataProperty: exp.Properties.Location.POST_EXPRESSION, 261 exp.WithJournalTableProperty: exp.Properties.Location.POST_NAME, 262 } 263 264 # Keywords that can't be used as unquoted identifier names 265 RESERVED_KEYWORDS: t.Set[str] = set() 266 267 # Expressions whose comments are separated from them for better formatting 268 WITH_SEPARATED_COMMENTS: t.Tuple[t.Type[exp.Expression], ...] = ( 269 exp.Drop, 270 exp.From, 271 exp.Select, 272 exp.Where, 273 exp.With, 274 ) 275 276 # Expressions that can remain unwrapped when appearing in the context of an INTERVAL 277 UNWRAPPED_INTERVAL_VALUES: t.Tuple[t.Type[exp.Expression], ...] = ( 278 exp.Column, 279 exp.Literal, 280 exp.Neg, 281 exp.Paren, 282 ) 283 284 SENTINEL_LINE_BREAK = "__SQLGLOT__LB__" 285 286 # Autofilled 287 INVERSE_TIME_MAPPING: t.Dict[str, str] = {} 288 INVERSE_TIME_TRIE: t.Dict = {} 289 INDEX_OFFSET = 0 290 UNNEST_COLUMN_ONLY = False 291 ALIAS_POST_TABLESAMPLE = False 292 IDENTIFIERS_CAN_START_WITH_DIGIT = False 293 STRICT_STRING_CONCAT = False 294 NORMALIZE_FUNCTIONS: bool | str = "upper" 295 NULL_ORDERING = "nulls_are_small" 296 ESCAPE_LINE_BREAK = False 297 298 can_identify: t.Callable[[str, str | bool], bool] 299 300 # Delimiters for quotes, identifiers and the corresponding escape characters 301 QUOTE_START = "'" 302 QUOTE_END = "'" 303 IDENTIFIER_START = '"' 304 IDENTIFIER_END = '"' 305 STRING_ESCAPE = "'" 306 IDENTIFIER_ESCAPE = '"' 307 308 # Delimiters for bit, hex, byte and raw literals 309 BIT_START: t.Optional[str] = None 310 BIT_END: t.Optional[str] = None 311 HEX_START: t.Optional[str] = None 312 HEX_END: t.Optional[str] = None 313 BYTE_START: t.Optional[str] = None 314 BYTE_END: t.Optional[str] = None 315 316 __slots__ = ( 317 "pretty", 318 "identify", 319 "normalize", 320 "pad", 321 "_indent", 322 "normalize_functions", 323 "unsupported_level", 324 "max_unsupported", 325 "leading_comma", 326 "max_text_width", 327 "comments", 328 "unsupported_messages", 329 "_escaped_quote_end", 330 "_escaped_identifier_end", 331 "_cache", 332 ) 333 334 def __init__( 335 self, 336 pretty: t.Optional[bool] = None, 337 identify: str | bool = False, 338 normalize: bool = False, 339 pad: int = 2, 340 indent: int = 2, 341 normalize_functions: t.Optional[str | bool] = None, 342 unsupported_level: ErrorLevel = ErrorLevel.WARN, 343 max_unsupported: int = 3, 344 leading_comma: bool = False, 345 max_text_width: int = 80, 346 comments: bool = True, 347 ): 348 import sqlglot 349 350 self.pretty = pretty if pretty is not None else sqlglot.pretty 351 self.identify = identify 352 self.normalize = normalize 353 self.pad = pad 354 self._indent = indent 355 self.unsupported_level = unsupported_level 356 self.max_unsupported = max_unsupported 357 self.leading_comma = leading_comma 358 self.max_text_width = max_text_width 359 self.comments = comments 360 361 # This is both a Dialect property and a Generator argument, so we prioritize the latter 362 self.normalize_functions = ( 363 self.NORMALIZE_FUNCTIONS if normalize_functions is None else normalize_functions 364 ) 365 366 self.unsupported_messages: t.List[str] = [] 367 self._escaped_quote_end: str = self.STRING_ESCAPE + self.QUOTE_END 368 self._escaped_identifier_end: str = self.IDENTIFIER_ESCAPE + self.IDENTIFIER_END 369 self._cache: t.Optional[t.Dict[int, str]] = None 370 371 def generate( 372 self, 373 expression: t.Optional[exp.Expression], 374 cache: t.Optional[t.Dict[int, str]] = None, 375 ) -> str: 376 """ 377 Generates the SQL string corresponding to the given syntax tree. 378 379 Args: 380 expression: The syntax tree. 381 cache: An optional sql string cache. This leverages the hash of an Expression 382 which can be slow to compute, so only use it if you set _hash on each node. 383 384 Returns: 385 The SQL string corresponding to `expression`. 386 """ 387 if cache is not None: 388 self._cache = cache 389 390 self.unsupported_messages = [] 391 sql = self.sql(expression).strip() 392 self._cache = None 393 394 if self.unsupported_level == ErrorLevel.IGNORE: 395 return sql 396 397 if self.unsupported_level == ErrorLevel.WARN: 398 for msg in self.unsupported_messages: 399 logger.warning(msg) 400 elif self.unsupported_level == ErrorLevel.RAISE and self.unsupported_messages: 401 raise UnsupportedError(concat_messages(self.unsupported_messages, self.max_unsupported)) 402 403 if self.pretty: 404 sql = sql.replace(self.SENTINEL_LINE_BREAK, "\n") 405 return sql 406 407 def unsupported(self, message: str) -> None: 408 if self.unsupported_level == ErrorLevel.IMMEDIATE: 409 raise UnsupportedError(message) 410 self.unsupported_messages.append(message) 411 412 def sep(self, sep: str = " ") -> str: 413 return f"{sep.strip()}\n" if self.pretty else sep 414 415 def seg(self, sql: str, sep: str = " ") -> str: 416 return f"{self.sep(sep)}{sql}" 417 418 def pad_comment(self, comment: str) -> str: 419 comment = " " + comment if comment[0].strip() else comment 420 comment = comment + " " if comment[-1].strip() else comment 421 return comment 422 423 def maybe_comment( 424 self, 425 sql: str, 426 expression: t.Optional[exp.Expression] = None, 427 comments: t.Optional[t.List[str]] = None, 428 ) -> str: 429 comments = ( 430 ((expression and expression.comments) if comments is None else comments) # type: ignore 431 if self.comments 432 else None 433 ) 434 435 if not comments or isinstance(expression, exp.Binary): 436 return sql 437 438 sep = "\n" if self.pretty else " " 439 comments_sql = sep.join( 440 f"/*{self.pad_comment(comment)}*/" for comment in comments if comment 441 ) 442 443 if not comments_sql: 444 return sql 445 446 if isinstance(expression, self.WITH_SEPARATED_COMMENTS): 447 return ( 448 f"{self.sep()}{comments_sql}{sql}" 449 if sql[0].isspace() 450 else f"{comments_sql}{self.sep()}{sql}" 451 ) 452 453 return f"{sql} {comments_sql}" 454 455 def wrap(self, expression: exp.Expression | str) -> str: 456 this_sql = self.indent( 457 self.sql(expression) 458 if isinstance(expression, (exp.Select, exp.Union)) 459 else self.sql(expression, "this"), 460 level=1, 461 pad=0, 462 ) 463 return f"({self.sep('')}{this_sql}{self.seg(')', sep='')}" 464 465 def no_identify(self, func: t.Callable[..., str], *args, **kwargs) -> str: 466 original = self.identify 467 self.identify = False 468 result = func(*args, **kwargs) 469 self.identify = original 470 return result 471 472 def normalize_func(self, name: str) -> str: 473 if self.normalize_functions == "upper" or self.normalize_functions is True: 474 return name.upper() 475 if self.normalize_functions == "lower": 476 return name.lower() 477 return name 478 479 def indent( 480 self, 481 sql: str, 482 level: int = 0, 483 pad: t.Optional[int] = None, 484 skip_first: bool = False, 485 skip_last: bool = False, 486 ) -> str: 487 if not self.pretty: 488 return sql 489 490 pad = self.pad if pad is None else pad 491 lines = sql.split("\n") 492 493 return "\n".join( 494 line 495 if (skip_first and i == 0) or (skip_last and i == len(lines) - 1) 496 else f"{' ' * (level * self._indent + pad)}{line}" 497 for i, line in enumerate(lines) 498 ) 499 500 def sql( 501 self, 502 expression: t.Optional[str | exp.Expression], 503 key: t.Optional[str] = None, 504 comment: bool = True, 505 ) -> str: 506 if not expression: 507 return "" 508 509 if isinstance(expression, str): 510 return expression 511 512 if key: 513 value = expression.args.get(key) 514 if value: 515 return self.sql(value) 516 return "" 517 518 if self._cache is not None: 519 expression_id = hash(expression) 520 521 if expression_id in self._cache: 522 return self._cache[expression_id] 523 524 transform = self.TRANSFORMS.get(expression.__class__) 525 526 if callable(transform): 527 sql = transform(self, expression) 528 elif transform: 529 sql = transform 530 elif isinstance(expression, exp.Expression): 531 exp_handler_name = f"{expression.key}_sql" 532 533 if hasattr(self, exp_handler_name): 534 sql = getattr(self, exp_handler_name)(expression) 535 elif isinstance(expression, exp.Func): 536 sql = self.function_fallback_sql(expression) 537 elif isinstance(expression, exp.Property): 538 sql = self.property_sql(expression) 539 else: 540 raise ValueError(f"Unsupported expression type {expression.__class__.__name__}") 541 else: 542 raise ValueError(f"Expected an Expression. Received {type(expression)}: {expression}") 543 544 sql = self.maybe_comment(sql, expression) if self.comments and comment else sql 545 546 if self._cache is not None: 547 self._cache[expression_id] = sql 548 return sql 549 550 def uncache_sql(self, expression: exp.Uncache) -> str: 551 table = self.sql(expression, "this") 552 exists_sql = " IF EXISTS" if expression.args.get("exists") else "" 553 return f"UNCACHE TABLE{exists_sql} {table}" 554 555 def cache_sql(self, expression: exp.Cache) -> str: 556 lazy = " LAZY" if expression.args.get("lazy") else "" 557 table = self.sql(expression, "this") 558 options = expression.args.get("options") 559 options = f" OPTIONS({self.sql(options[0])} = {self.sql(options[1])})" if options else "" 560 sql = self.sql(expression, "expression") 561 sql = f" AS{self.sep()}{sql}" if sql else "" 562 sql = f"CACHE{lazy} TABLE {table}{options}{sql}" 563 return self.prepend_ctes(expression, sql) 564 565 def characterset_sql(self, expression: exp.CharacterSet) -> str: 566 if isinstance(expression.parent, exp.Cast): 567 return f"CHAR CHARACTER SET {self.sql(expression, 'this')}" 568 default = "DEFAULT " if expression.args.get("default") else "" 569 return f"{default}CHARACTER SET={self.sql(expression, 'this')}" 570 571 def column_sql(self, expression: exp.Column) -> str: 572 join_mark = " (+)" if expression.args.get("join_mark") else "" 573 574 if join_mark and not self.COLUMN_JOIN_MARKS_SUPPORTED: 575 join_mark = "" 576 self.unsupported("Outer join syntax using the (+) operator is not supported.") 577 578 column = ".".join( 579 self.sql(part) 580 for part in ( 581 expression.args.get("catalog"), 582 expression.args.get("db"), 583 expression.args.get("table"), 584 expression.args.get("this"), 585 ) 586 if part 587 ) 588 589 return f"{column}{join_mark}" 590 591 def columnposition_sql(self, expression: exp.ColumnPosition) -> str: 592 this = self.sql(expression, "this") 593 this = f" {this}" if this else "" 594 position = self.sql(expression, "position") 595 return f"{position}{this}" 596 597 def columndef_sql(self, expression: exp.ColumnDef, sep: str = " ") -> str: 598 column = self.sql(expression, "this") 599 kind = self.sql(expression, "kind") 600 constraints = self.expressions(expression, key="constraints", sep=" ", flat=True) 601 exists = "IF NOT EXISTS " if expression.args.get("exists") else "" 602 kind = f"{sep}{kind}" if kind else "" 603 constraints = f" {constraints}" if constraints else "" 604 position = self.sql(expression, "position") 605 position = f" {position}" if position else "" 606 607 return f"{exists}{column}{kind}{constraints}{position}" 608 609 def columnconstraint_sql(self, expression: exp.ColumnConstraint) -> str: 610 this = self.sql(expression, "this") 611 kind_sql = self.sql(expression, "kind").strip() 612 return f"CONSTRAINT {this} {kind_sql}" if this else kind_sql 613 614 def autoincrementcolumnconstraint_sql(self, _) -> str: 615 return self.token_sql(TokenType.AUTO_INCREMENT) 616 617 def compresscolumnconstraint_sql(self, expression: exp.CompressColumnConstraint) -> str: 618 if isinstance(expression.this, list): 619 this = self.wrap(self.expressions(expression, key="this", flat=True)) 620 else: 621 this = self.sql(expression, "this") 622 623 return f"COMPRESS {this}" 624 625 def generatedasidentitycolumnconstraint_sql( 626 self, expression: exp.GeneratedAsIdentityColumnConstraint 627 ) -> str: 628 this = "" 629 if expression.this is not None: 630 on_null = "ON NULL " if expression.args.get("on_null") else "" 631 this = " ALWAYS " if expression.this else f" BY DEFAULT {on_null}" 632 633 start = expression.args.get("start") 634 start = f"START WITH {start}" if start else "" 635 increment = expression.args.get("increment") 636 increment = f" INCREMENT BY {increment}" if increment else "" 637 minvalue = expression.args.get("minvalue") 638 minvalue = f" MINVALUE {minvalue}" if minvalue else "" 639 maxvalue = expression.args.get("maxvalue") 640 maxvalue = f" MAXVALUE {maxvalue}" if maxvalue else "" 641 cycle = expression.args.get("cycle") 642 cycle_sql = "" 643 644 if cycle is not None: 645 cycle_sql = f"{' NO' if not cycle else ''} CYCLE" 646 cycle_sql = cycle_sql.strip() if not start and not increment else cycle_sql 647 648 sequence_opts = "" 649 if start or increment or cycle_sql: 650 sequence_opts = f"{start}{increment}{minvalue}{maxvalue}{cycle_sql}" 651 sequence_opts = f" ({sequence_opts.strip()})" 652 653 expr = self.sql(expression, "expression") 654 expr = f"({expr})" if expr else "IDENTITY" 655 656 return f"GENERATED{this}AS {expr}{sequence_opts}" 657 658 def notnullcolumnconstraint_sql(self, expression: exp.NotNullColumnConstraint) -> str: 659 return f"{'' if expression.args.get('allow_null') else 'NOT '}NULL" 660 661 def primarykeycolumnconstraint_sql(self, expression: exp.PrimaryKeyColumnConstraint) -> str: 662 desc = expression.args.get("desc") 663 if desc is not None: 664 return f"PRIMARY KEY{' DESC' if desc else ' ASC'}" 665 return f"PRIMARY KEY" 666 667 def uniquecolumnconstraint_sql(self, expression: exp.UniqueColumnConstraint) -> str: 668 this = self.sql(expression, "this") 669 this = f" {this}" if this else "" 670 return f"UNIQUE{this}" 671 672 def createable_sql( 673 self, expression: exp.Create, locations: dict[exp.Properties.Location, list[exp.Property]] 674 ) -> str: 675 return self.sql(expression, "this") 676 677 def create_sql(self, expression: exp.Create) -> str: 678 kind = self.sql(expression, "kind").upper() 679 properties = expression.args.get("properties") 680 properties_locs = self.locate_properties(properties) if properties else {} 681 682 this = self.createable_sql(expression, properties_locs) 683 684 properties_sql = "" 685 if properties_locs.get(exp.Properties.Location.POST_SCHEMA) or properties_locs.get( 686 exp.Properties.Location.POST_WITH 687 ): 688 properties_sql = self.sql( 689 exp.Properties( 690 expressions=[ 691 *properties_locs[exp.Properties.Location.POST_SCHEMA], 692 *properties_locs[exp.Properties.Location.POST_WITH], 693 ] 694 ) 695 ) 696 697 begin = " BEGIN" if expression.args.get("begin") else "" 698 expression_sql = self.sql(expression, "expression") 699 if expression_sql: 700 expression_sql = f"{begin}{self.sep()}{expression_sql}" 701 702 if self.CREATE_FUNCTION_RETURN_AS or not isinstance(expression.expression, exp.Return): 703 if properties_locs.get(exp.Properties.Location.POST_ALIAS): 704 postalias_props_sql = self.properties( 705 exp.Properties( 706 expressions=properties_locs[exp.Properties.Location.POST_ALIAS] 707 ), 708 wrapped=False, 709 ) 710 expression_sql = f" AS {postalias_props_sql}{expression_sql}" 711 else: 712 expression_sql = f" AS{expression_sql}" 713 714 postindex_props_sql = "" 715 if properties_locs.get(exp.Properties.Location.POST_INDEX): 716 postindex_props_sql = self.properties( 717 exp.Properties(expressions=properties_locs[exp.Properties.Location.POST_INDEX]), 718 wrapped=False, 719 prefix=" ", 720 ) 721 722 indexes = self.expressions(expression, key="indexes", indent=False, sep=" ") 723 indexes = f" {indexes}" if indexes else "" 724 index_sql = indexes + postindex_props_sql 725 726 replace = " OR REPLACE" if expression.args.get("replace") else "" 727 unique = " UNIQUE" if expression.args.get("unique") else "" 728 729 postcreate_props_sql = "" 730 if properties_locs.get(exp.Properties.Location.POST_CREATE): 731 postcreate_props_sql = self.properties( 732 exp.Properties(expressions=properties_locs[exp.Properties.Location.POST_CREATE]), 733 sep=" ", 734 prefix=" ", 735 wrapped=False, 736 ) 737 738 modifiers = "".join((replace, unique, postcreate_props_sql)) 739 740 postexpression_props_sql = "" 741 if properties_locs.get(exp.Properties.Location.POST_EXPRESSION): 742 postexpression_props_sql = self.properties( 743 exp.Properties( 744 expressions=properties_locs[exp.Properties.Location.POST_EXPRESSION] 745 ), 746 sep=" ", 747 prefix=" ", 748 wrapped=False, 749 ) 750 751 exists_sql = " IF NOT EXISTS" if expression.args.get("exists") else "" 752 no_schema_binding = ( 753 " WITH NO SCHEMA BINDING" if expression.args.get("no_schema_binding") else "" 754 ) 755 756 clone = self.sql(expression, "clone") 757 clone = f" {clone}" if clone else "" 758 759 expression_sql = f"CREATE{modifiers} {kind}{exists_sql} {this}{properties_sql}{expression_sql}{postexpression_props_sql}{index_sql}{no_schema_binding}{clone}" 760 return self.prepend_ctes(expression, expression_sql) 761 762 def clone_sql(self, expression: exp.Clone) -> str: 763 this = self.sql(expression, "this") 764 when = self.sql(expression, "when") 765 766 if when: 767 kind = self.sql(expression, "kind") 768 expr = self.sql(expression, "expression") 769 return f"CLONE {this} {when} ({kind} => {expr})" 770 771 return f"CLONE {this}" 772 773 def describe_sql(self, expression: exp.Describe) -> str: 774 return f"DESCRIBE {self.sql(expression, 'this')}" 775 776 def prepend_ctes(self, expression: exp.Expression, sql: str) -> str: 777 with_ = self.sql(expression, "with") 778 if with_: 779 sql = f"{with_}{self.sep()}{sql}" 780 return sql 781 782 def with_sql(self, expression: exp.With) -> str: 783 sql = self.expressions(expression, flat=True) 784 recursive = "RECURSIVE " if expression.args.get("recursive") else "" 785 786 return f"WITH {recursive}{sql}" 787 788 def cte_sql(self, expression: exp.CTE) -> str: 789 alias = self.sql(expression, "alias") 790 return f"{alias} AS {self.wrap(expression)}" 791 792 def tablealias_sql(self, expression: exp.TableAlias) -> str: 793 alias = self.sql(expression, "this") 794 columns = self.expressions(expression, key="columns", flat=True) 795 columns = f"({columns})" if columns else "" 796 return f"{alias}{columns}" 797 798 def bitstring_sql(self, expression: exp.BitString) -> str: 799 this = self.sql(expression, "this") 800 if self.BIT_START: 801 return f"{self.BIT_START}{this}{self.BIT_END}" 802 return f"{int(this, 2)}" 803 804 def hexstring_sql(self, expression: exp.HexString) -> str: 805 this = self.sql(expression, "this") 806 if self.HEX_START: 807 return f"{self.HEX_START}{this}{self.HEX_END}" 808 return f"{int(this, 16)}" 809 810 def bytestring_sql(self, expression: exp.ByteString) -> str: 811 this = self.sql(expression, "this") 812 if self.BYTE_START: 813 return f"{self.BYTE_START}{this}{self.BYTE_END}" 814 return this 815 816 def rawstring_sql(self, expression: exp.RawString) -> str: 817 string = self.escape_str(expression.this.replace("\\", "\\\\")) 818 return f"{self.QUOTE_START}{string}{self.QUOTE_END}" 819 820 def datatypesize_sql(self, expression: exp.DataTypeSize) -> str: 821 this = self.sql(expression, "this") 822 specifier = self.sql(expression, "expression") 823 specifier = f" {specifier}" if specifier else "" 824 return f"{this}{specifier}" 825 826 def datatype_sql(self, expression: exp.DataType) -> str: 827 type_value = expression.this 828 type_sql = ( 829 self.TYPE_MAPPING.get(type_value, type_value.value) 830 if isinstance(type_value, exp.DataType.Type) 831 else type_value 832 ) 833 nested = "" 834 interior = self.expressions(expression, flat=True) 835 values = "" 836 if interior: 837 if expression.args.get("nested"): 838 nested = f"{self.STRUCT_DELIMITER[0]}{interior}{self.STRUCT_DELIMITER[1]}" 839 if expression.args.get("values") is not None: 840 delimiters = ("[", "]") if type_value == exp.DataType.Type.ARRAY else ("(", ")") 841 values = self.expressions(expression, key="values", flat=True) 842 values = f"{delimiters[0]}{values}{delimiters[1]}" 843 else: 844 nested = f"({interior})" 845 846 return f"{type_sql}{nested}{values}" 847 848 def directory_sql(self, expression: exp.Directory) -> str: 849 local = "LOCAL " if expression.args.get("local") else "" 850 row_format = self.sql(expression, "row_format") 851 row_format = f" {row_format}" if row_format else "" 852 return f"{local}DIRECTORY {self.sql(expression, 'this')}{row_format}" 853 854 def delete_sql(self, expression: exp.Delete) -> str: 855 this = self.sql(expression, "this") 856 this = f" FROM {this}" if this else "" 857 using = self.sql(expression, "using") 858 using = f" USING {using}" if using else "" 859 where = self.sql(expression, "where") 860 returning = self.sql(expression, "returning") 861 limit = self.sql(expression, "limit") 862 tables = self.expressions(expression, key="tables") 863 tables = f" {tables}" if tables else "" 864 if self.RETURNING_END: 865 expression_sql = f"{this}{using}{where}{returning}{limit}" 866 else: 867 expression_sql = f"{returning}{this}{using}{where}{limit}" 868 return self.prepend_ctes(expression, f"DELETE{tables}{expression_sql}") 869 870 def drop_sql(self, expression: exp.Drop) -> str: 871 this = self.sql(expression, "this") 872 kind = expression.args["kind"] 873 exists_sql = " IF EXISTS " if expression.args.get("exists") else " " 874 temporary = " TEMPORARY" if expression.args.get("temporary") else "" 875 materialized = " MATERIALIZED" if expression.args.get("materialized") else "" 876 cascade = " CASCADE" if expression.args.get("cascade") else "" 877 constraints = " CONSTRAINTS" if expression.args.get("constraints") else "" 878 purge = " PURGE" if expression.args.get("purge") else "" 879 return ( 880 f"DROP{temporary}{materialized} {kind}{exists_sql}{this}{cascade}{constraints}{purge}" 881 ) 882 883 def except_sql(self, expression: exp.Except) -> str: 884 return self.prepend_ctes( 885 expression, 886 self.set_operation(expression, self.except_op(expression)), 887 ) 888 889 def except_op(self, expression: exp.Except) -> str: 890 return f"EXCEPT{'' if expression.args.get('distinct') else ' ALL'}" 891 892 def fetch_sql(self, expression: exp.Fetch) -> str: 893 direction = expression.args.get("direction") 894 direction = f" {direction.upper()}" if direction else "" 895 count = expression.args.get("count") 896 count = f" {count}" if count else "" 897 if expression.args.get("percent"): 898 count = f"{count} PERCENT" 899 with_ties_or_only = "WITH TIES" if expression.args.get("with_ties") else "ONLY" 900 return f"{self.seg('FETCH')}{direction}{count} ROWS {with_ties_or_only}" 901 902 def filter_sql(self, expression: exp.Filter) -> str: 903 this = self.sql(expression, "this") 904 where = self.sql(expression, "expression")[1:] # where has a leading space 905 return f"{this} FILTER({where})" 906 907 def hint_sql(self, expression: exp.Hint) -> str: 908 if not self.QUERY_HINTS: 909 self.unsupported("Hints are not supported") 910 return "" 911 912 return f" /*+ {self.expressions(expression, sep=self.QUERY_HINT_SEP).strip()} */" 913 914 def index_sql(self, expression: exp.Index) -> str: 915 unique = "UNIQUE " if expression.args.get("unique") else "" 916 primary = "PRIMARY " if expression.args.get("primary") else "" 917 amp = "AMP " if expression.args.get("amp") else "" 918 name = self.sql(expression, "this") 919 name = f"{name} " if name else "" 920 table = self.sql(expression, "table") 921 table = f"{self.INDEX_ON} {table} " if table else "" 922 using = self.sql(expression, "using") 923 using = f"USING {using} " if using else "" 924 index = "INDEX " if not table else "" 925 columns = self.expressions(expression, key="columns", flat=True) 926 columns = f"({columns})" if columns else "" 927 partition_by = self.expressions(expression, key="partition_by", flat=True) 928 partition_by = f" PARTITION BY {partition_by}" if partition_by else "" 929 return f"{unique}{primary}{amp}{index}{name}{table}{using}{columns}{partition_by}" 930 931 def identifier_sql(self, expression: exp.Identifier) -> str: 932 text = expression.name 933 lower = text.lower() 934 text = lower if self.normalize and not expression.quoted else text 935 text = text.replace(self.IDENTIFIER_END, self._escaped_identifier_end) 936 if ( 937 expression.quoted 938 or self.can_identify(text, self.identify) 939 or lower in self.RESERVED_KEYWORDS 940 or (not self.IDENTIFIERS_CAN_START_WITH_DIGIT and text[:1].isdigit()) 941 ): 942 text = f"{self.IDENTIFIER_START}{text}{self.IDENTIFIER_END}" 943 return text 944 945 def inputoutputformat_sql(self, expression: exp.InputOutputFormat) -> str: 946 input_format = self.sql(expression, "input_format") 947 input_format = f"INPUTFORMAT {input_format}" if input_format else "" 948 output_format = self.sql(expression, "output_format") 949 output_format = f"OUTPUTFORMAT {output_format}" if output_format else "" 950 return self.sep().join((input_format, output_format)) 951 952 def national_sql(self, expression: exp.National, prefix: str = "N") -> str: 953 string = self.sql(exp.Literal.string(expression.name)) 954 return f"{prefix}{string}" 955 956 def partition_sql(self, expression: exp.Partition) -> str: 957 return f"PARTITION({self.expressions(expression, flat=True)})" 958 959 def properties_sql(self, expression: exp.Properties) -> str: 960 root_properties = [] 961 with_properties = [] 962 963 for p in expression.expressions: 964 p_loc = self.PROPERTIES_LOCATION[p.__class__] 965 if p_loc == exp.Properties.Location.POST_WITH: 966 with_properties.append(p) 967 elif p_loc == exp.Properties.Location.POST_SCHEMA: 968 root_properties.append(p) 969 970 return self.root_properties( 971 exp.Properties(expressions=root_properties) 972 ) + self.with_properties(exp.Properties(expressions=with_properties)) 973 974 def root_properties(self, properties: exp.Properties) -> str: 975 if properties.expressions: 976 return self.sep() + self.expressions(properties, indent=False, sep=" ") 977 return "" 978 979 def properties( 980 self, 981 properties: exp.Properties, 982 prefix: str = "", 983 sep: str = ", ", 984 suffix: str = "", 985 wrapped: bool = True, 986 ) -> str: 987 if properties.expressions: 988 expressions = self.expressions(properties, sep=sep, indent=False) 989 expressions = self.wrap(expressions) if wrapped else expressions 990 return f"{prefix}{' ' if prefix and prefix != ' ' else ''}{expressions}{suffix}" 991 return "" 992 993 def with_properties(self, properties: exp.Properties) -> str: 994 return self.properties(properties, prefix=self.seg("WITH")) 995 996 def locate_properties( 997 self, properties: exp.Properties 998 ) -> t.Dict[exp.Properties.Location, list[exp.Property]]: 999 properties_locs: t.Dict[exp.Properties.Location, list[exp.Property]] = { 1000 key: [] for key in exp.Properties.Location 1001 } 1002 1003 for p in properties.expressions: 1004 p_loc = self.PROPERTIES_LOCATION[p.__class__] 1005 if p_loc == exp.Properties.Location.POST_NAME: 1006 properties_locs[exp.Properties.Location.POST_NAME].append(p) 1007 elif p_loc == exp.Properties.Location.POST_INDEX: 1008 properties_locs[exp.Properties.Location.POST_INDEX].append(p) 1009 elif p_loc == exp.Properties.Location.POST_SCHEMA: 1010 properties_locs[exp.Properties.Location.POST_SCHEMA].append(p) 1011 elif p_loc == exp.Properties.Location.POST_WITH: 1012 properties_locs[exp.Properties.Location.POST_WITH].append(p) 1013 elif p_loc == exp.Properties.Location.POST_CREATE: 1014 properties_locs[exp.Properties.Location.POST_CREATE].append(p) 1015 elif p_loc == exp.Properties.Location.POST_ALIAS: 1016 properties_locs[exp.Properties.Location.POST_ALIAS].append(p) 1017 elif p_loc == exp.Properties.Location.POST_EXPRESSION: 1018 properties_locs[exp.Properties.Location.POST_EXPRESSION].append(p) 1019 elif p_loc == exp.Properties.Location.UNSUPPORTED: 1020 self.unsupported(f"Unsupported property {p.key}") 1021 1022 return properties_locs 1023 1024 def property_sql(self, expression: exp.Property) -> str: 1025 property_cls = expression.__class__ 1026 if property_cls == exp.Property: 1027 return f"{expression.name}={self.sql(expression, 'value')}" 1028 1029 property_name = exp.Properties.PROPERTY_TO_NAME.get(property_cls) 1030 if not property_name: 1031 self.unsupported(f"Unsupported property {expression.key}") 1032 1033 return f"{property_name}={self.sql(expression, 'this')}" 1034 1035 def likeproperty_sql(self, expression: exp.LikeProperty) -> str: 1036 options = " ".join(f"{e.name} {self.sql(e, 'value')}" for e in expression.expressions) 1037 options = f" {options}" if options else "" 1038 return f"LIKE {self.sql(expression, 'this')}{options}" 1039 1040 def fallbackproperty_sql(self, expression: exp.FallbackProperty) -> str: 1041 no = "NO " if expression.args.get("no") else "" 1042 protection = " PROTECTION" if expression.args.get("protection") else "" 1043 return f"{no}FALLBACK{protection}" 1044 1045 def journalproperty_sql(self, expression: exp.JournalProperty) -> str: 1046 no = "NO " if expression.args.get("no") else "" 1047 local = expression.args.get("local") 1048 local = f"{local} " if local else "" 1049 dual = "DUAL " if expression.args.get("dual") else "" 1050 before = "BEFORE " if expression.args.get("before") else "" 1051 after = "AFTER " if expression.args.get("after") else "" 1052 return f"{no}{local}{dual}{before}{after}JOURNAL" 1053 1054 def freespaceproperty_sql(self, expression: exp.FreespaceProperty) -> str: 1055 freespace = self.sql(expression, "this") 1056 percent = " PERCENT" if expression.args.get("percent") else "" 1057 return f"FREESPACE={freespace}{percent}" 1058 1059 def checksumproperty_sql(self, expression: exp.ChecksumProperty) -> str: 1060 if expression.args.get("default"): 1061 property = "DEFAULT" 1062 elif expression.args.get("on"): 1063 property = "ON" 1064 else: 1065 property = "OFF" 1066 return f"CHECKSUM={property}" 1067 1068 def mergeblockratioproperty_sql(self, expression: exp.MergeBlockRatioProperty) -> str: 1069 if expression.args.get("no"): 1070 return "NO MERGEBLOCKRATIO" 1071 if expression.args.get("default"): 1072 return "DEFAULT MERGEBLOCKRATIO" 1073 1074 percent = " PERCENT" if expression.args.get("percent") else "" 1075 return f"MERGEBLOCKRATIO={self.sql(expression, 'this')}{percent}" 1076 1077 def datablocksizeproperty_sql(self, expression: exp.DataBlocksizeProperty) -> str: 1078 default = expression.args.get("default") 1079 minimum = expression.args.get("minimum") 1080 maximum = expression.args.get("maximum") 1081 if default or minimum or maximum: 1082 if default: 1083 prop = "DEFAULT" 1084 elif minimum: 1085 prop = "MINIMUM" 1086 else: 1087 prop = "MAXIMUM" 1088 return f"{prop} DATABLOCKSIZE" 1089 units = expression.args.get("units") 1090 units = f" {units}" if units else "" 1091 return f"DATABLOCKSIZE={self.sql(expression, 'size')}{units}" 1092 1093 def blockcompressionproperty_sql(self, expression: exp.BlockCompressionProperty) -> str: 1094 autotemp = expression.args.get("autotemp") 1095 always = expression.args.get("always") 1096 default = expression.args.get("default") 1097 manual = expression.args.get("manual") 1098 never = expression.args.get("never") 1099 1100 if autotemp is not None: 1101 prop = f"AUTOTEMP({self.expressions(autotemp)})" 1102 elif always: 1103 prop = "ALWAYS" 1104 elif default: 1105 prop = "DEFAULT" 1106 elif manual: 1107 prop = "MANUAL" 1108 elif never: 1109 prop = "NEVER" 1110 return f"BLOCKCOMPRESSION={prop}" 1111 1112 def isolatedloadingproperty_sql(self, expression: exp.IsolatedLoadingProperty) -> str: 1113 no = expression.args.get("no") 1114 no = " NO" if no else "" 1115 concurrent = expression.args.get("concurrent") 1116 concurrent = " CONCURRENT" if concurrent else "" 1117 1118 for_ = "" 1119 if expression.args.get("for_all"): 1120 for_ = " FOR ALL" 1121 elif expression.args.get("for_insert"): 1122 for_ = " FOR INSERT" 1123 elif expression.args.get("for_none"): 1124 for_ = " FOR NONE" 1125 return f"WITH{no}{concurrent} ISOLATED LOADING{for_}" 1126 1127 def lockingproperty_sql(self, expression: exp.LockingProperty) -> str: 1128 kind = expression.args.get("kind") 1129 this = f" {self.sql(expression, 'this')}" if expression.this else "" 1130 for_or_in = expression.args.get("for_or_in") 1131 lock_type = expression.args.get("lock_type") 1132 override = " OVERRIDE" if expression.args.get("override") else "" 1133 return f"LOCKING {kind}{this} {for_or_in} {lock_type}{override}" 1134 1135 def withdataproperty_sql(self, expression: exp.WithDataProperty) -> str: 1136 data_sql = f"WITH {'NO ' if expression.args.get('no') else ''}DATA" 1137 statistics = expression.args.get("statistics") 1138 statistics_sql = "" 1139 if statistics is not None: 1140 statistics_sql = f" AND {'NO ' if not statistics else ''}STATISTICS" 1141 return f"{data_sql}{statistics_sql}" 1142 1143 def insert_sql(self, expression: exp.Insert) -> str: 1144 overwrite = expression.args.get("overwrite") 1145 1146 if isinstance(expression.this, exp.Directory): 1147 this = " OVERWRITE" if overwrite else " INTO" 1148 else: 1149 this = " OVERWRITE TABLE" if overwrite else " INTO" 1150 1151 alternative = expression.args.get("alternative") 1152 alternative = f" OR {alternative}" if alternative else "" 1153 ignore = " IGNORE" if expression.args.get("ignore") else "" 1154 1155 this = f"{this} {self.sql(expression, 'this')}" 1156 1157 exists = " IF EXISTS" if expression.args.get("exists") else "" 1158 partition_sql = ( 1159 f" {self.sql(expression, 'partition')}" if expression.args.get("partition") else "" 1160 ) 1161 where = self.sql(expression, "where") 1162 where = f"{self.sep()}REPLACE WHERE {where}" if where else "" 1163 expression_sql = f"{self.sep()}{self.sql(expression, 'expression')}" 1164 conflict = self.sql(expression, "conflict") 1165 returning = self.sql(expression, "returning") 1166 1167 if self.RETURNING_END: 1168 expression_sql = f"{expression_sql}{conflict}{returning}" 1169 else: 1170 expression_sql = f"{returning}{expression_sql}{conflict}" 1171 1172 sql = f"INSERT{alternative}{ignore}{this}{exists}{partition_sql}{where}{expression_sql}" 1173 return self.prepend_ctes(expression, sql) 1174 1175 def intersect_sql(self, expression: exp.Intersect) -> str: 1176 return self.prepend_ctes( 1177 expression, 1178 self.set_operation(expression, self.intersect_op(expression)), 1179 ) 1180 1181 def intersect_op(self, expression: exp.Intersect) -> str: 1182 return f"INTERSECT{'' if expression.args.get('distinct') else ' ALL'}" 1183 1184 def introducer_sql(self, expression: exp.Introducer) -> str: 1185 return f"{self.sql(expression, 'this')} {self.sql(expression, 'expression')}" 1186 1187 def pseudotype_sql(self, expression: exp.PseudoType) -> str: 1188 return expression.name.upper() 1189 1190 def onconflict_sql(self, expression: exp.OnConflict) -> str: 1191 conflict = "ON DUPLICATE KEY" if expression.args.get("duplicate") else "ON CONFLICT" 1192 constraint = self.sql(expression, "constraint") 1193 if constraint: 1194 constraint = f"ON CONSTRAINT {constraint}" 1195 key = self.expressions(expression, key="key", flat=True) 1196 do = "" if expression.args.get("duplicate") else " DO " 1197 nothing = "NOTHING" if expression.args.get("nothing") else "" 1198 expressions = self.expressions(expression, flat=True) 1199 set_keyword = "SET " if self.DUPLICATE_KEY_UPDATE_WITH_SET else "" 1200 if expressions: 1201 expressions = f"UPDATE {set_keyword}{expressions}" 1202 return f"{self.seg(conflict)} {constraint}{key}{do}{nothing}{expressions}" 1203 1204 def returning_sql(self, expression: exp.Returning) -> str: 1205 return f"{self.seg('RETURNING')} {self.expressions(expression, flat=True)}" 1206 1207 def rowformatdelimitedproperty_sql(self, expression: exp.RowFormatDelimitedProperty) -> str: 1208 fields = expression.args.get("fields") 1209 fields = f" FIELDS TERMINATED BY {fields}" if fields else "" 1210 escaped = expression.args.get("escaped") 1211 escaped = f" ESCAPED BY {escaped}" if escaped else "" 1212 items = expression.args.get("collection_items") 1213 items = f" COLLECTION ITEMS TERMINATED BY {items}" if items else "" 1214 keys = expression.args.get("map_keys") 1215 keys = f" MAP KEYS TERMINATED BY {keys}" if keys else "" 1216 lines = expression.args.get("lines") 1217 lines = f" LINES TERMINATED BY {lines}" if lines else "" 1218 null = expression.args.get("null") 1219 null = f" NULL DEFINED AS {null}" if null else "" 1220 return f"ROW FORMAT DELIMITED{fields}{escaped}{items}{keys}{lines}{null}" 1221 1222 def withtablehint_sql(self, expression: exp.WithTableHint) -> str: 1223 return f"WITH ({self.expressions(expression, flat=True)})" 1224 1225 def indextablehint_sql(self, expression: exp.IndexTableHint) -> str: 1226 this = f"{self.sql(expression, 'this')} INDEX" 1227 target = self.sql(expression, "target") 1228 target = f" FOR {target}" if target else "" 1229 return f"{this}{target} ({self.expressions(expression, flat=True)})" 1230 1231 def table_sql(self, expression: exp.Table, sep: str = " AS ") -> str: 1232 table = ".".join( 1233 part 1234 for part in [ 1235 self.sql(expression, "catalog"), 1236 self.sql(expression, "db"), 1237 self.sql(expression, "this"), 1238 ] 1239 if part 1240 ) 1241 1242 alias = self.sql(expression, "alias") 1243 alias = f"{sep}{alias}" if alias else "" 1244 hints = self.expressions(expression, key="hints", sep=" ") 1245 hints = f" {hints}" if hints and self.TABLE_HINTS else "" 1246 pivots = self.expressions(expression, key="pivots", sep=" ", flat=True) 1247 pivots = f" {pivots}" if pivots else "" 1248 joins = self.expressions(expression, key="joins", sep="", skip_first=True) 1249 laterals = self.expressions(expression, key="laterals", sep="") 1250 system_time = expression.args.get("system_time") 1251 system_time = f" {self.sql(expression, 'system_time')}" if system_time else "" 1252 1253 return f"{table}{system_time}{alias}{hints}{pivots}{joins}{laterals}" 1254 1255 def tablesample_sql( 1256 self, expression: exp.TableSample, seed_prefix: str = "SEED", sep=" AS " 1257 ) -> str: 1258 if self.ALIAS_POST_TABLESAMPLE and expression.this.alias: 1259 table = expression.this.copy() 1260 table.set("alias", None) 1261 this = self.sql(table) 1262 alias = f"{sep}{self.sql(expression.this, 'alias')}" 1263 else: 1264 this = self.sql(expression, "this") 1265 alias = "" 1266 method = self.sql(expression, "method") 1267 method = f"{method.upper()} " if method and self.TABLESAMPLE_WITH_METHOD else "" 1268 numerator = self.sql(expression, "bucket_numerator") 1269 denominator = self.sql(expression, "bucket_denominator") 1270 field = self.sql(expression, "bucket_field") 1271 field = f" ON {field}" if field else "" 1272 bucket = f"BUCKET {numerator} OUT OF {denominator}{field}" if numerator else "" 1273 percent = self.sql(expression, "percent") 1274 percent = f"{percent} PERCENT" if percent else "" 1275 rows = self.sql(expression, "rows") 1276 rows = f"{rows} ROWS" if rows else "" 1277 size = self.sql(expression, "size") 1278 if size and self.TABLESAMPLE_SIZE_IS_PERCENT: 1279 size = f"{size} PERCENT" 1280 seed = self.sql(expression, "seed") 1281 seed = f" {seed_prefix} ({seed})" if seed else "" 1282 kind = expression.args.get("kind", "TABLESAMPLE") 1283 return f"{this} {kind} {method}({bucket}{percent}{rows}{size}){seed}{alias}" 1284 1285 def pivot_sql(self, expression: exp.Pivot) -> str: 1286 expressions = self.expressions(expression, flat=True) 1287 1288 if expression.this: 1289 this = self.sql(expression, "this") 1290 on = f"{self.seg('ON')} {expressions}" 1291 using = self.expressions(expression, key="using", flat=True) 1292 using = f"{self.seg('USING')} {using}" if using else "" 1293 group = self.sql(expression, "group") 1294 return f"PIVOT {this}{on}{using}{group}" 1295 1296 alias = self.sql(expression, "alias") 1297 alias = f" AS {alias}" if alias else "" 1298 unpivot = expression.args.get("unpivot") 1299 direction = "UNPIVOT" if unpivot else "PIVOT" 1300 field = self.sql(expression, "field") 1301 return f"{direction}({expressions} FOR {field}){alias}" 1302 1303 def tuple_sql(self, expression: exp.Tuple) -> str: 1304 return f"({self.expressions(expression, flat=True)})" 1305 1306 def update_sql(self, expression: exp.Update) -> str: 1307 this = self.sql(expression, "this") 1308 set_sql = self.expressions(expression, flat=True) 1309 from_sql = self.sql(expression, "from") 1310 where_sql = self.sql(expression, "where") 1311 returning = self.sql(expression, "returning") 1312 limit = self.sql(expression, "limit") 1313 if self.RETURNING_END: 1314 expression_sql = f"{from_sql}{where_sql}{returning}{limit}" 1315 else: 1316 expression_sql = f"{returning}{from_sql}{where_sql}{limit}" 1317 sql = f"UPDATE {this} SET {set_sql}{expression_sql}" 1318 return self.prepend_ctes(expression, sql) 1319 1320 def values_sql(self, expression: exp.Values) -> str: 1321 # The VALUES clause is still valid in an `INSERT INTO ..` statement, for example 1322 if self.VALUES_AS_TABLE or not expression.find_ancestor(exp.From, exp.Join): 1323 args = self.expressions(expression) 1324 alias = self.sql(expression, "alias") 1325 values = f"VALUES{self.seg('')}{args}" 1326 values = ( 1327 f"({values})" 1328 if self.WRAP_DERIVED_VALUES and (alias or isinstance(expression.parent, exp.From)) 1329 else values 1330 ) 1331 return f"{values} AS {alias}" if alias else values 1332 1333 # Converts `VALUES...` expression into a series of select unions. 1334 # Note: If you have a lot of unions then this will result in a large number of recursive statements to 1335 # evaluate the expression. You may need to increase `sys.setrecursionlimit` to run and it can also be 1336 # very slow. 1337 expression = expression.copy() 1338 column_names = expression.alias and expression.args["alias"].columns 1339 1340 selects = [] 1341 1342 for i, tup in enumerate(expression.expressions): 1343 row = tup.expressions 1344 1345 if i == 0 and column_names: 1346 row = [ 1347 exp.alias_(value, column_name) for value, column_name in zip(row, column_names) 1348 ] 1349 1350 selects.append(exp.Select(expressions=row)) 1351 1352 subquery_expression: exp.Select | exp.Union = selects[0] 1353 if len(selects) > 1: 1354 for select in selects[1:]: 1355 subquery_expression = exp.union( 1356 subquery_expression, select, distinct=False, copy=False 1357 ) 1358 1359 return self.subquery_sql(subquery_expression.subquery(expression.alias, copy=False)) 1360 1361 def var_sql(self, expression: exp.Var) -> str: 1362 return self.sql(expression, "this") 1363 1364 def into_sql(self, expression: exp.Into) -> str: 1365 temporary = " TEMPORARY" if expression.args.get("temporary") else "" 1366 unlogged = " UNLOGGED" if expression.args.get("unlogged") else "" 1367 return f"{self.seg('INTO')}{temporary or unlogged} {self.sql(expression, 'this')}" 1368 1369 def from_sql(self, expression: exp.From) -> str: 1370 return f"{self.seg('FROM')} {self.sql(expression, 'this')}" 1371 1372 def group_sql(self, expression: exp.Group) -> str: 1373 group_by = self.op_expressions("GROUP BY", expression) 1374 1375 if expression.args.get("all"): 1376 return f"{group_by} ALL" 1377 1378 grouping_sets = self.expressions(expression, key="grouping_sets", indent=False) 1379 grouping_sets = ( 1380 f"{self.seg('GROUPING SETS')} {self.wrap(grouping_sets)}" if grouping_sets else "" 1381 ) 1382 1383 cube = expression.args.get("cube", []) 1384 if seq_get(cube, 0) is True: 1385 return f"{group_by}{self.seg('WITH CUBE')}" 1386 else: 1387 cube_sql = self.expressions(expression, key="cube", indent=False) 1388 cube_sql = f"{self.seg('CUBE')} {self.wrap(cube_sql)}" if cube_sql else "" 1389 1390 rollup = expression.args.get("rollup", []) 1391 if seq_get(rollup, 0) is True: 1392 return f"{group_by}{self.seg('WITH ROLLUP')}" 1393 else: 1394 rollup_sql = self.expressions(expression, key="rollup", indent=False) 1395 rollup_sql = f"{self.seg('ROLLUP')} {self.wrap(rollup_sql)}" if rollup_sql else "" 1396 1397 groupings = csv( 1398 grouping_sets, 1399 cube_sql, 1400 rollup_sql, 1401 self.seg("WITH TOTALS") if expression.args.get("totals") else "", 1402 sep=self.GROUPINGS_SEP, 1403 ) 1404 1405 if expression.args.get("expressions") and groupings: 1406 group_by = f"{group_by}{self.GROUPINGS_SEP}" 1407 1408 return f"{group_by}{groupings}" 1409 1410 def having_sql(self, expression: exp.Having) -> str: 1411 this = self.indent(self.sql(expression, "this")) 1412 return f"{self.seg('HAVING')}{self.sep()}{this}" 1413 1414 def join_sql(self, expression: exp.Join) -> str: 1415 op_sql = " ".join( 1416 op 1417 for op in ( 1418 expression.method, 1419 "GLOBAL" if expression.args.get("global") else None, 1420 expression.side, 1421 expression.kind, 1422 expression.hint if self.JOIN_HINTS else None, 1423 ) 1424 if op 1425 ) 1426 on_sql = self.sql(expression, "on") 1427 using = expression.args.get("using") 1428 1429 if not on_sql and using: 1430 on_sql = csv(*(self.sql(column) for column in using)) 1431 1432 this_sql = self.sql(expression, "this") 1433 1434 if on_sql: 1435 on_sql = self.indent(on_sql, skip_first=True) 1436 space = self.seg(" " * self.pad) if self.pretty else " " 1437 if using: 1438 on_sql = f"{space}USING ({on_sql})" 1439 else: 1440 on_sql = f"{space}ON {on_sql}" 1441 elif not op_sql: 1442 return f", {this_sql}" 1443 1444 op_sql = f"{op_sql} JOIN" if op_sql else "JOIN" 1445 return f"{self.seg(op_sql)} {this_sql}{on_sql}" 1446 1447 def lambda_sql(self, expression: exp.Lambda, arrow_sep: str = "->") -> str: 1448 args = self.expressions(expression, flat=True) 1449 args = f"({args})" if len(args.split(",")) > 1 else args 1450 return f"{args} {arrow_sep} {self.sql(expression, 'this')}" 1451 1452 def lateral_sql(self, expression: exp.Lateral) -> str: 1453 this = self.sql(expression, "this") 1454 1455 if isinstance(expression.this, exp.Subquery): 1456 return f"LATERAL {this}" 1457 1458 if expression.args.get("view"): 1459 alias = expression.args["alias"] 1460 columns = self.expressions(alias, key="columns", flat=True) 1461 table = f" {alias.name}" if alias.name else "" 1462 columns = f" AS {columns}" if columns else "" 1463 op_sql = self.seg(f"LATERAL VIEW{' OUTER' if expression.args.get('outer') else ''}") 1464 return f"{op_sql}{self.sep()}{this}{table}{columns}" 1465 1466 alias = self.sql(expression, "alias") 1467 alias = f" AS {alias}" if alias else "" 1468 return f"LATERAL {this}{alias}" 1469 1470 def limit_sql(self, expression: exp.Limit, top: bool = False) -> str: 1471 this = self.sql(expression, "this") 1472 args = ", ".join( 1473 sql 1474 for sql in ( 1475 self.sql(expression, "offset"), 1476 self.sql(expression, "expression"), 1477 ) 1478 if sql 1479 ) 1480 return f"{this}{self.seg('TOP' if top else 'LIMIT')} {args}" 1481 1482 def offset_sql(self, expression: exp.Offset) -> str: 1483 this = self.sql(expression, "this") 1484 return f"{this}{self.seg('OFFSET')} {self.sql(expression, 'expression')}" 1485 1486 def setitem_sql(self, expression: exp.SetItem) -> str: 1487 kind = self.sql(expression, "kind") 1488 kind = f"{kind} " if kind else "" 1489 this = self.sql(expression, "this") 1490 expressions = self.expressions(expression) 1491 collate = self.sql(expression, "collate") 1492 collate = f" COLLATE {collate}" if collate else "" 1493 global_ = "GLOBAL " if expression.args.get("global") else "" 1494 return f"{global_}{kind}{this}{expressions}{collate}" 1495 1496 def set_sql(self, expression: exp.Set) -> str: 1497 expressions = ( 1498 f" {self.expressions(expression, flat=True)}" if expression.expressions else "" 1499 ) 1500 tag = " TAG" if expression.args.get("tag") else "" 1501 return f"{'UNSET' if expression.args.get('unset') else 'SET'}{tag}{expressions}" 1502 1503 def pragma_sql(self, expression: exp.Pragma) -> str: 1504 return f"PRAGMA {self.sql(expression, 'this')}" 1505 1506 def lock_sql(self, expression: exp.Lock) -> str: 1507 if not self.LOCKING_READS_SUPPORTED: 1508 self.unsupported("Locking reads using 'FOR UPDATE/SHARE' are not supported") 1509 return "" 1510 1511 lock_type = "FOR UPDATE" if expression.args["update"] else "FOR SHARE" 1512 expressions = self.expressions(expression, flat=True) 1513 expressions = f" OF {expressions}" if expressions else "" 1514 wait = expression.args.get("wait") 1515 1516 if wait is not None: 1517 if isinstance(wait, exp.Literal): 1518 wait = f" WAIT {self.sql(wait)}" 1519 else: 1520 wait = " NOWAIT" if wait else " SKIP LOCKED" 1521 1522 return f"{lock_type}{expressions}{wait or ''}" 1523 1524 def literal_sql(self, expression: exp.Literal) -> str: 1525 text = expression.this or "" 1526 if expression.is_string: 1527 text = f"{self.QUOTE_START}{self.escape_str(text)}{self.QUOTE_END}" 1528 return text 1529 1530 def escape_str(self, text: str) -> str: 1531 text = text.replace(self.QUOTE_END, self._escaped_quote_end) 1532 if self.ESCAPE_LINE_BREAK: 1533 text = text.replace("\n", "\\n") 1534 elif self.pretty: 1535 text = text.replace("\n", self.SENTINEL_LINE_BREAK) 1536 return text 1537 1538 def loaddata_sql(self, expression: exp.LoadData) -> str: 1539 local = " LOCAL" if expression.args.get("local") else "" 1540 inpath = f" INPATH {self.sql(expression, 'inpath')}" 1541 overwrite = " OVERWRITE" if expression.args.get("overwrite") else "" 1542 this = f" INTO TABLE {self.sql(expression, 'this')}" 1543 partition = self.sql(expression, "partition") 1544 partition = f" {partition}" if partition else "" 1545 input_format = self.sql(expression, "input_format") 1546 input_format = f" INPUTFORMAT {input_format}" if input_format else "" 1547 serde = self.sql(expression, "serde") 1548 serde = f" SERDE {serde}" if serde else "" 1549 return f"LOAD DATA{local}{inpath}{overwrite}{this}{partition}{input_format}{serde}" 1550 1551 def null_sql(self, *_) -> str: 1552 return "NULL" 1553 1554 def boolean_sql(self, expression: exp.Boolean) -> str: 1555 return "TRUE" if expression.this else "FALSE" 1556 1557 def order_sql(self, expression: exp.Order, flat: bool = False) -> str: 1558 this = self.sql(expression, "this") 1559 this = f"{this} " if this else this 1560 return self.op_expressions(f"{this}ORDER BY", expression, flat=this or flat) # type: ignore 1561 1562 def cluster_sql(self, expression: exp.Cluster) -> str: 1563 return self.op_expressions("CLUSTER BY", expression) 1564 1565 def distribute_sql(self, expression: exp.Distribute) -> str: 1566 return self.op_expressions("DISTRIBUTE BY", expression) 1567 1568 def sort_sql(self, expression: exp.Sort) -> str: 1569 return self.op_expressions("SORT BY", expression) 1570 1571 def ordered_sql(self, expression: exp.Ordered) -> str: 1572 desc = expression.args.get("desc") 1573 asc = not desc 1574 1575 nulls_first = expression.args.get("nulls_first") 1576 nulls_last = not nulls_first 1577 nulls_are_large = self.NULL_ORDERING == "nulls_are_large" 1578 nulls_are_small = self.NULL_ORDERING == "nulls_are_small" 1579 nulls_are_last = self.NULL_ORDERING == "nulls_are_last" 1580 1581 sort_order = " DESC" if desc else "" 1582 nulls_sort_change = "" 1583 if nulls_first and ( 1584 (asc and nulls_are_large) or (desc and nulls_are_small) or nulls_are_last 1585 ): 1586 nulls_sort_change = " NULLS FIRST" 1587 elif ( 1588 nulls_last 1589 and ((asc and nulls_are_small) or (desc and nulls_are_large)) 1590 and not nulls_are_last 1591 ): 1592 nulls_sort_change = " NULLS LAST" 1593 1594 if nulls_sort_change and not self.NULL_ORDERING_SUPPORTED: 1595 self.unsupported( 1596 "Sorting in an ORDER BY on NULLS FIRST/NULLS LAST is not supported by this dialect" 1597 ) 1598 nulls_sort_change = "" 1599 1600 return f"{self.sql(expression, 'this')}{sort_order}{nulls_sort_change}" 1601 1602 def matchrecognize_sql(self, expression: exp.MatchRecognize) -> str: 1603 partition = self.partition_by_sql(expression) 1604 order = self.sql(expression, "order") 1605 measures = self.expressions(expression, key="measures") 1606 measures = self.seg(f"MEASURES{self.seg(measures)}") if measures else "" 1607 rows = self.sql(expression, "rows") 1608 rows = self.seg(rows) if rows else "" 1609 after = self.sql(expression, "after") 1610 after = self.seg(after) if after else "" 1611 pattern = self.sql(expression, "pattern") 1612 pattern = self.seg(f"PATTERN ({pattern})") if pattern else "" 1613 definition_sqls = [ 1614 f"{self.sql(definition, 'alias')} AS {self.sql(definition, 'this')}" 1615 for definition in expression.args.get("define", []) 1616 ] 1617 definitions = self.expressions(sqls=definition_sqls) 1618 define = self.seg(f"DEFINE{self.seg(definitions)}") if definitions else "" 1619 body = "".join( 1620 ( 1621 partition, 1622 order, 1623 measures, 1624 rows, 1625 after, 1626 pattern, 1627 define, 1628 ) 1629 ) 1630 alias = self.sql(expression, "alias") 1631 alias = f" {alias}" if alias else "" 1632 return f"{self.seg('MATCH_RECOGNIZE')} {self.wrap(body)}{alias}" 1633 1634 def query_modifiers(self, expression: exp.Expression, *sqls: str) -> str: 1635 limit: t.Optional[exp.Fetch | exp.Limit] = expression.args.get("limit") 1636 1637 # If the limit is generated as TOP, we need to ensure it's not generated twice 1638 with_offset_limit_modifiers = not isinstance(limit, exp.Limit) or not self.LIMIT_IS_TOP 1639 1640 if self.LIMIT_FETCH == "LIMIT" and isinstance(limit, exp.Fetch): 1641 limit = exp.Limit(expression=limit.args.get("count")) 1642 elif self.LIMIT_FETCH == "FETCH" and isinstance(limit, exp.Limit): 1643 limit = exp.Fetch(direction="FIRST", count=limit.expression) 1644 1645 fetch = isinstance(limit, exp.Fetch) 1646 1647 offset_limit_modifiers = ( 1648 self.offset_limit_modifiers(expression, fetch, limit) 1649 if with_offset_limit_modifiers 1650 else [] 1651 ) 1652 1653 return csv( 1654 *sqls, 1655 *[self.sql(join) for join in expression.args.get("joins") or []], 1656 self.sql(expression, "match"), 1657 *[self.sql(lateral) for lateral in expression.args.get("laterals") or []], 1658 self.sql(expression, "where"), 1659 self.sql(expression, "group"), 1660 self.sql(expression, "having"), 1661 *self.after_having_modifiers(expression), 1662 self.sql(expression, "order"), 1663 *offset_limit_modifiers, 1664 *self.after_limit_modifiers(expression), 1665 sep="", 1666 ) 1667 1668 def offset_limit_modifiers( 1669 self, expression: exp.Expression, fetch: bool, limit: t.Optional[exp.Fetch | exp.Limit] 1670 ) -> t.List[str]: 1671 return [ 1672 self.sql(expression, "offset") if fetch else self.sql(limit), 1673 self.sql(limit) if fetch else self.sql(expression, "offset"), 1674 ] 1675 1676 def after_having_modifiers(self, expression: exp.Expression) -> t.List[str]: 1677 return [ 1678 self.sql(expression, "qualify"), 1679 self.seg("WINDOW ") + self.expressions(expression, key="windows", flat=True) 1680 if expression.args.get("windows") 1681 else "", 1682 self.sql(expression, "distribute"), 1683 self.sql(expression, "sort"), 1684 self.sql(expression, "cluster"), 1685 ] 1686 1687 def after_limit_modifiers(self, expression: exp.Expression) -> t.List[str]: 1688 locks = self.expressions(expression, key="locks", sep=" ") 1689 locks = f" {locks}" if locks else "" 1690 return [locks, self.sql(expression, "sample")] 1691 1692 def select_sql(self, expression: exp.Select) -> str: 1693 hint = self.sql(expression, "hint") 1694 distinct = self.sql(expression, "distinct") 1695 distinct = f" {distinct}" if distinct else "" 1696 kind = self.sql(expression, "kind").upper() 1697 limit = expression.args.get("limit") 1698 top = ( 1699 self.limit_sql(limit, top=True) 1700 if isinstance(limit, exp.Limit) and self.LIMIT_IS_TOP 1701 else "" 1702 ) 1703 1704 expressions = self.expressions(expression) 1705 1706 if kind: 1707 if kind in self.SELECT_KINDS: 1708 kind = f" AS {kind}" 1709 else: 1710 if kind == "STRUCT": 1711 expressions = self.expressions( 1712 sqls=[ 1713 self.sql( 1714 exp.Struct( 1715 expressions=[ 1716 exp.column(e.output_name).eq( 1717 e.this if isinstance(e, exp.Alias) else e 1718 ) 1719 for e in expression.expressions 1720 ] 1721 ) 1722 ) 1723 ] 1724 ) 1725 kind = "" 1726 1727 expressions = f"{self.sep()}{expressions}" if expressions else expressions 1728 sql = self.query_modifiers( 1729 expression, 1730 f"SELECT{top}{hint}{distinct}{kind}{expressions}", 1731 self.sql(expression, "into", comment=False), 1732 self.sql(expression, "from", comment=False), 1733 ) 1734 return self.prepend_ctes(expression, sql) 1735 1736 def schema_sql(self, expression: exp.Schema) -> str: 1737 this = self.sql(expression, "this") 1738 this = f"{this} " if this else "" 1739 sql = self.schema_columns_sql(expression) 1740 return f"{this}{sql}" 1741 1742 def schema_columns_sql(self, expression: exp.Schema) -> str: 1743 return f"({self.sep('')}{self.expressions(expression)}{self.seg(')', sep='')}" 1744 1745 def star_sql(self, expression: exp.Star) -> str: 1746 except_ = self.expressions(expression, key="except", flat=True) 1747 except_ = f"{self.seg(self.STAR_MAPPING['except'])} ({except_})" if except_ else "" 1748 replace = self.expressions(expression, key="replace", flat=True) 1749 replace = f"{self.seg(self.STAR_MAPPING['replace'])} ({replace})" if replace else "" 1750 return f"*{except_}{replace}" 1751 1752 def parameter_sql(self, expression: exp.Parameter) -> str: 1753 this = self.sql(expression, "this") 1754 this = f"{{{this}}}" if expression.args.get("wrapped") else f"{this}" 1755 return f"{self.PARAMETER_TOKEN}{this}" 1756 1757 def sessionparameter_sql(self, expression: exp.SessionParameter) -> str: 1758 this = self.sql(expression, "this") 1759 kind = expression.text("kind") 1760 if kind: 1761 kind = f"{kind}." 1762 return f"@@{kind}{this}" 1763 1764 def placeholder_sql(self, expression: exp.Placeholder) -> str: 1765 return f":{expression.name}" if expression.name else "?" 1766 1767 def subquery_sql(self, expression: exp.Subquery, sep: str = " AS ") -> str: 1768 alias = self.sql(expression, "alias") 1769 alias = f"{sep}{alias}" if alias else "" 1770 1771 pivots = self.expressions(expression, key="pivots", sep=" ", flat=True) 1772 pivots = f" {pivots}" if pivots else "" 1773 1774 sql = self.query_modifiers(expression, self.wrap(expression), alias, pivots) 1775 return self.prepend_ctes(expression, sql) 1776 1777 def qualify_sql(self, expression: exp.Qualify) -> str: 1778 this = self.indent(self.sql(expression, "this")) 1779 return f"{self.seg('QUALIFY')}{self.sep()}{this}" 1780 1781 def union_sql(self, expression: exp.Union) -> str: 1782 return self.prepend_ctes( 1783 expression, 1784 self.set_operation(expression, self.union_op(expression)), 1785 ) 1786 1787 def union_op(self, expression: exp.Union) -> str: 1788 kind = " DISTINCT" if self.EXPLICIT_UNION else "" 1789 kind = kind if expression.args.get("distinct") else " ALL" 1790 return f"UNION{kind}" 1791 1792 def unnest_sql(self, expression: exp.Unnest) -> str: 1793 args = self.expressions(expression, flat=True) 1794 alias = expression.args.get("alias") 1795 if alias and self.UNNEST_COLUMN_ONLY: 1796 columns = alias.columns 1797 alias = self.sql(columns[0]) if columns else "" 1798 else: 1799 alias = self.sql(expression, "alias") 1800 alias = f" AS {alias}" if alias else alias 1801 ordinality = " WITH ORDINALITY" if expression.args.get("ordinality") else "" 1802 offset = expression.args.get("offset") 1803 offset = f" WITH OFFSET AS {self.sql(offset)}" if offset else "" 1804 return f"UNNEST({args}){ordinality}{alias}{offset}" 1805 1806 def where_sql(self, expression: exp.Where) -> str: 1807 this = self.indent(self.sql(expression, "this")) 1808 return f"{self.seg('WHERE')}{self.sep()}{this}" 1809 1810 def window_sql(self, expression: exp.Window) -> str: 1811 this = self.sql(expression, "this") 1812 partition = self.partition_by_sql(expression) 1813 order = expression.args.get("order") 1814 order = self.order_sql(order, flat=True) if order else "" 1815 spec = self.sql(expression, "spec") 1816 alias = self.sql(expression, "alias") 1817 over = self.sql(expression, "over") or "OVER" 1818 1819 this = f"{this} {'AS' if expression.arg_key == 'windows' else over}" 1820 1821 first = expression.args.get("first") 1822 if first is None: 1823 first = "" 1824 else: 1825 first = "FIRST" if first else "LAST" 1826 1827 if not partition and not order and not spec and alias: 1828 return f"{this} {alias}" 1829 1830 args = " ".join(arg for arg in (alias, first, partition, order, spec) if arg) 1831 return f"{this} ({args})" 1832 1833 def partition_by_sql(self, expression: exp.Window | exp.MatchRecognize) -> str: 1834 partition = self.expressions(expression, key="partition_by", flat=True) 1835 return f"PARTITION BY {partition}" if partition else "" 1836 1837 def windowspec_sql(self, expression: exp.WindowSpec) -> str: 1838 kind = self.sql(expression, "kind") 1839 start = csv(self.sql(expression, "start"), self.sql(expression, "start_side"), sep=" ") 1840 end = ( 1841 csv(self.sql(expression, "end"), self.sql(expression, "end_side"), sep=" ") 1842 or "CURRENT ROW" 1843 ) 1844 return f"{kind} BETWEEN {start} AND {end}" 1845 1846 def withingroup_sql(self, expression: exp.WithinGroup) -> str: 1847 this = self.sql(expression, "this") 1848 expression_sql = self.sql(expression, "expression")[1:] # order has a leading space 1849 return f"{this} WITHIN GROUP ({expression_sql})" 1850 1851 def between_sql(self, expression: exp.Between) -> str: 1852 this = self.sql(expression, "this") 1853 low = self.sql(expression, "low") 1854 high = self.sql(expression, "high") 1855 return f"{this} BETWEEN {low} AND {high}" 1856 1857 def bracket_sql(self, expression: exp.Bracket) -> str: 1858 expressions = apply_index_offset(expression.this, expression.expressions, self.INDEX_OFFSET) 1859 expressions_sql = ", ".join(self.sql(e) for e in expressions) 1860 1861 return f"{self.sql(expression, 'this')}[{expressions_sql}]" 1862 1863 def safebracket_sql(self, expression: exp.SafeBracket) -> str: 1864 return self.bracket_sql(expression) 1865 1866 def all_sql(self, expression: exp.All) -> str: 1867 return f"ALL {self.wrap(expression)}" 1868 1869 def any_sql(self, expression: exp.Any) -> str: 1870 this = self.sql(expression, "this") 1871 if isinstance(expression.this, exp.Subqueryable): 1872 this = self.wrap(this) 1873 return f"ANY {this}" 1874 1875 def exists_sql(self, expression: exp.Exists) -> str: 1876 return f"EXISTS{self.wrap(expression)}" 1877 1878 def case_sql(self, expression: exp.Case) -> str: 1879 this = self.sql(expression, "this") 1880 statements = [f"CASE {this}" if this else "CASE"] 1881 1882 for e in expression.args["ifs"]: 1883 statements.append(f"WHEN {self.sql(e, 'this')}") 1884 statements.append(f"THEN {self.sql(e, 'true')}") 1885 1886 default = self.sql(expression, "default") 1887 1888 if default: 1889 statements.append(f"ELSE {default}") 1890 1891 statements.append("END") 1892 1893 if self.pretty and self.text_width(statements) > self.max_text_width: 1894 return self.indent("\n".join(statements), skip_first=True, skip_last=True) 1895 1896 return " ".join(statements) 1897 1898 def constraint_sql(self, expression: exp.Constraint) -> str: 1899 this = self.sql(expression, "this") 1900 expressions = self.expressions(expression, flat=True) 1901 return f"CONSTRAINT {this} {expressions}" 1902 1903 def nextvaluefor_sql(self, expression: exp.NextValueFor) -> str: 1904 order = expression.args.get("order") 1905 order = f" OVER ({self.order_sql(order, flat=True)})" if order else "" 1906 return f"NEXT VALUE FOR {self.sql(expression, 'this')}{order}" 1907 1908 def extract_sql(self, expression: exp.Extract) -> str: 1909 this = self.sql(expression, "this") 1910 expression_sql = self.sql(expression, "expression") 1911 return f"EXTRACT({this} FROM {expression_sql})" 1912 1913 def trim_sql(self, expression: exp.Trim) -> str: 1914 trim_type = self.sql(expression, "position") 1915 1916 if trim_type == "LEADING": 1917 return self.func("LTRIM", expression.this) 1918 elif trim_type == "TRAILING": 1919 return self.func("RTRIM", expression.this) 1920 else: 1921 return self.func("TRIM", expression.this, expression.expression) 1922 1923 def safeconcat_sql(self, expression: exp.SafeConcat) -> str: 1924 expressions = expression.expressions 1925 if self.STRICT_STRING_CONCAT: 1926 expressions = (exp.cast(e, "text") for e in expressions) 1927 return self.func("CONCAT", *expressions) 1928 1929 def check_sql(self, expression: exp.Check) -> str: 1930 this = self.sql(expression, key="this") 1931 return f"CHECK ({this})" 1932 1933 def foreignkey_sql(self, expression: exp.ForeignKey) -> str: 1934 expressions = self.expressions(expression, flat=True) 1935 reference = self.sql(expression, "reference") 1936 reference = f" {reference}" if reference else "" 1937 delete = self.sql(expression, "delete") 1938 delete = f" ON DELETE {delete}" if delete else "" 1939 update = self.sql(expression, "update") 1940 update = f" ON UPDATE {update}" if update else "" 1941 return f"FOREIGN KEY ({expressions}){reference}{delete}{update}" 1942 1943 def primarykey_sql(self, expression: exp.ForeignKey) -> str: 1944 expressions = self.expressions(expression, flat=True) 1945 options = self.expressions(expression, key="options", flat=True, sep=" ") 1946 options = f" {options}" if options else "" 1947 return f"PRIMARY KEY ({expressions}){options}" 1948 1949 def if_sql(self, expression: exp.If) -> str: 1950 return self.case_sql(exp.Case(ifs=[expression], default=expression.args.get("false"))) 1951 1952 def matchagainst_sql(self, expression: exp.MatchAgainst) -> str: 1953 modifier = expression.args.get("modifier") 1954 modifier = f" {modifier}" if modifier else "" 1955 return f"{self.func('MATCH', *expression.expressions)} AGAINST({self.sql(expression, 'this')}{modifier})" 1956 1957 def jsonkeyvalue_sql(self, expression: exp.JSONKeyValue) -> str: 1958 return f"{self.sql(expression, 'this')}: {self.sql(expression, 'expression')}" 1959 1960 def jsonobject_sql(self, expression: exp.JSONObject) -> str: 1961 null_handling = expression.args.get("null_handling") 1962 null_handling = f" {null_handling}" if null_handling else "" 1963 unique_keys = expression.args.get("unique_keys") 1964 if unique_keys is not None: 1965 unique_keys = f" {'WITH' if unique_keys else 'WITHOUT'} UNIQUE KEYS" 1966 else: 1967 unique_keys = "" 1968 return_type = self.sql(expression, "return_type") 1969 return_type = f" RETURNING {return_type}" if return_type else "" 1970 format_json = " FORMAT JSON" if expression.args.get("format_json") else "" 1971 encoding = self.sql(expression, "encoding") 1972 encoding = f" ENCODING {encoding}" if encoding else "" 1973 return self.func( 1974 "JSON_OBJECT", 1975 *expression.expressions, 1976 suffix=f"{null_handling}{unique_keys}{return_type}{format_json}{encoding})", 1977 ) 1978 1979 def openjsoncolumndef_sql(self, expression: exp.OpenJSONColumnDef) -> str: 1980 this = self.sql(expression, "this") 1981 kind = self.sql(expression, "kind") 1982 path = self.sql(expression, "path") 1983 path = f" {path}" if path else "" 1984 as_json = " AS JSON" if expression.args.get("as_json") else "" 1985 return f"{this} {kind}{path}{as_json}" 1986 1987 def openjson_sql(self, expression: exp.OpenJSON) -> str: 1988 this = self.sql(expression, "this") 1989 path = self.sql(expression, "path") 1990 path = f", {path}" if path else "" 1991 expressions = self.expressions(expression) 1992 with_ = ( 1993 f" WITH ({self.seg(self.indent(expressions), sep='')}{self.seg(')', sep='')}" 1994 if expressions 1995 else "" 1996 ) 1997 return f"OPENJSON({this}{path}){with_}" 1998 1999 def in_sql(self, expression: exp.In) -> str: 2000 query = expression.args.get("query") 2001 unnest = expression.args.get("unnest") 2002 field = expression.args.get("field") 2003 is_global = " GLOBAL" if expression.args.get("is_global") else "" 2004 2005 if query: 2006 in_sql = self.wrap(query) 2007 elif unnest: 2008 in_sql = self.in_unnest_op(unnest) 2009 elif field: 2010 in_sql = self.sql(field) 2011 else: 2012 in_sql = f"({self.expressions(expression, flat=True)})" 2013 2014 return f"{self.sql(expression, 'this')}{is_global} IN {in_sql}" 2015 2016 def in_unnest_op(self, unnest: exp.Unnest) -> str: 2017 return f"(SELECT {self.sql(unnest)})" 2018 2019 def interval_sql(self, expression: exp.Interval) -> str: 2020 unit = self.sql(expression, "unit") 2021 if not self.INTERVAL_ALLOWS_PLURAL_FORM: 2022 unit = self.TIME_PART_SINGULARS.get(unit.lower(), unit) 2023 unit = f" {unit}" if unit else "" 2024 2025 if self.SINGLE_STRING_INTERVAL: 2026 this = expression.this.name if expression.this else "" 2027 return f"INTERVAL '{this}{unit}'" if this else f"INTERVAL{unit}" 2028 2029 this = self.sql(expression, "this") 2030 if this: 2031 unwrapped = isinstance(expression.this, self.UNWRAPPED_INTERVAL_VALUES) 2032 this = f" {this}" if unwrapped else f" ({this})" 2033 2034 return f"INTERVAL{this}{unit}" 2035 2036 def return_sql(self, expression: exp.Return) -> str: 2037 return f"RETURN {self.sql(expression, 'this')}" 2038 2039 def reference_sql(self, expression: exp.Reference) -> str: 2040 this = self.sql(expression, "this") 2041 expressions = self.expressions(expression, flat=True) 2042 expressions = f"({expressions})" if expressions else "" 2043 options = self.expressions(expression, key="options", flat=True, sep=" ") 2044 options = f" {options}" if options else "" 2045 return f"REFERENCES {this}{expressions}{options}" 2046 2047 def anonymous_sql(self, expression: exp.Anonymous) -> str: 2048 return self.func(expression.name, *expression.expressions) 2049 2050 def paren_sql(self, expression: exp.Paren) -> str: 2051 if isinstance(expression.unnest(), exp.Select): 2052 sql = self.wrap(expression) 2053 else: 2054 sql = self.seg(self.indent(self.sql(expression, "this")), sep="") 2055 sql = f"({sql}{self.seg(')', sep='')}" 2056 2057 return self.prepend_ctes(expression, sql) 2058 2059 def neg_sql(self, expression: exp.Neg) -> str: 2060 # This makes sure we don't convert "- - 5" to "--5", which is a comment 2061 this_sql = self.sql(expression, "this") 2062 sep = " " if this_sql[0] == "-" else "" 2063 return f"-{sep}{this_sql}" 2064 2065 def not_sql(self, expression: exp.Not) -> str: 2066 return f"NOT {self.sql(expression, 'this')}" 2067 2068 def alias_sql(self, expression: exp.Alias) -> str: 2069 alias = self.sql(expression, "alias") 2070 alias = f" AS {alias}" if alias else "" 2071 return f"{self.sql(expression, 'this')}{alias}" 2072 2073 def aliases_sql(self, expression: exp.Aliases) -> str: 2074 return f"{self.sql(expression, 'this')} AS ({self.expressions(expression, flat=True)})" 2075 2076 def attimezone_sql(self, expression: exp.AtTimeZone) -> str: 2077 this = self.sql(expression, "this") 2078 zone = self.sql(expression, "zone") 2079 return f"{this} AT TIME ZONE {zone}" 2080 2081 def add_sql(self, expression: exp.Add) -> str: 2082 return self.binary(expression, "+") 2083 2084 def and_sql(self, expression: exp.And) -> str: 2085 return self.connector_sql(expression, "AND") 2086 2087 def xor_sql(self, expression: exp.Xor) -> str: 2088 return self.connector_sql(expression, "XOR") 2089 2090 def connector_sql(self, expression: exp.Connector, op: str) -> str: 2091 if not self.pretty: 2092 return self.binary(expression, op) 2093 2094 sqls = tuple( 2095 self.maybe_comment(self.sql(e), e, e.parent.comments or []) if i != 1 else self.sql(e) 2096 for i, e in enumerate(expression.flatten(unnest=False)) 2097 ) 2098 2099 sep = "\n" if self.text_width(sqls) > self.max_text_width else " " 2100 return f"{sep}{op} ".join(sqls) 2101 2102 def bitwiseand_sql(self, expression: exp.BitwiseAnd) -> str: 2103 return self.binary(expression, "&") 2104 2105 def bitwiseleftshift_sql(self, expression: exp.BitwiseLeftShift) -> str: 2106 return self.binary(expression, "<<") 2107 2108 def bitwisenot_sql(self, expression: exp.BitwiseNot) -> str: 2109 return f"~{self.sql(expression, 'this')}" 2110 2111 def bitwiseor_sql(self, expression: exp.BitwiseOr) -> str: 2112 return self.binary(expression, "|") 2113 2114 def bitwiserightshift_sql(self, expression: exp.BitwiseRightShift) -> str: 2115 return self.binary(expression, ">>") 2116 2117 def bitwisexor_sql(self, expression: exp.BitwiseXor) -> str: 2118 return self.binary(expression, "^") 2119 2120 def cast_sql(self, expression: exp.Cast, safe_prefix: t.Optional[str] = None) -> str: 2121 format_sql = self.sql(expression, "format") 2122 format_sql = f" FORMAT {format_sql}" if format_sql else "" 2123 return f"{safe_prefix or ''}CAST({self.sql(expression, 'this')} AS {self.sql(expression, 'to')}{format_sql})" 2124 2125 def currentdate_sql(self, expression: exp.CurrentDate) -> str: 2126 zone = self.sql(expression, "this") 2127 return f"CURRENT_DATE({zone})" if zone else "CURRENT_DATE" 2128 2129 def collate_sql(self, expression: exp.Collate) -> str: 2130 return self.binary(expression, "COLLATE") 2131 2132 def command_sql(self, expression: exp.Command) -> str: 2133 return f"{self.sql(expression, 'this').upper()} {expression.text('expression').strip()}" 2134 2135 def comment_sql(self, expression: exp.Comment) -> str: 2136 this = self.sql(expression, "this") 2137 kind = expression.args["kind"] 2138 exists_sql = " IF EXISTS " if expression.args.get("exists") else " " 2139 expression_sql = self.sql(expression, "expression") 2140 return f"COMMENT{exists_sql}ON {kind} {this} IS {expression_sql}" 2141 2142 def mergetreettlaction_sql(self, expression: exp.MergeTreeTTLAction) -> str: 2143 this = self.sql(expression, "this") 2144 delete = " DELETE" if expression.args.get("delete") else "" 2145 recompress = self.sql(expression, "recompress") 2146 recompress = f" RECOMPRESS {recompress}" if recompress else "" 2147 to_disk = self.sql(expression, "to_disk") 2148 to_disk = f" TO DISK {to_disk}" if to_disk else "" 2149 to_volume = self.sql(expression, "to_volume") 2150 to_volume = f" TO VOLUME {to_volume}" if to_volume else "" 2151 return f"{this}{delete}{recompress}{to_disk}{to_volume}" 2152 2153 def mergetreettl_sql(self, expression: exp.MergeTreeTTL) -> str: 2154 where = self.sql(expression, "where") 2155 group = self.sql(expression, "group") 2156 aggregates = self.expressions(expression, key="aggregates") 2157 aggregates = self.seg("SET") + self.seg(aggregates) if aggregates else "" 2158 2159 if not (where or group or aggregates) and len(expression.expressions) == 1: 2160 return f"TTL {self.expressions(expression, flat=True)}" 2161 2162 return f"TTL{self.seg(self.expressions(expression))}{where}{group}{aggregates}" 2163 2164 def transaction_sql(self, expression: exp.Transaction) -> str: 2165 return "BEGIN" 2166 2167 def commit_sql(self, expression: exp.Commit) -> str: 2168 chain = expression.args.get("chain") 2169 if chain is not None: 2170 chain = " AND CHAIN" if chain else " AND NO CHAIN" 2171 2172 return f"COMMIT{chain or ''}" 2173 2174 def rollback_sql(self, expression: exp.Rollback) -> str: 2175 savepoint = expression.args.get("savepoint") 2176 savepoint = f" TO {savepoint}" if savepoint else "" 2177 return f"ROLLBACK{savepoint}" 2178 2179 def altercolumn_sql(self, expression: exp.AlterColumn) -> str: 2180 this = self.sql(expression, "this") 2181 2182 dtype = self.sql(expression, "dtype") 2183 if dtype: 2184 collate = self.sql(expression, "collate") 2185 collate = f" COLLATE {collate}" if collate else "" 2186 using = self.sql(expression, "using") 2187 using = f" USING {using}" if using else "" 2188 return f"ALTER COLUMN {this} TYPE {dtype}{collate}{using}" 2189 2190 default = self.sql(expression, "default") 2191 if default: 2192 return f"ALTER COLUMN {this} SET DEFAULT {default}" 2193 2194 if not expression.args.get("drop"): 2195 self.unsupported("Unsupported ALTER COLUMN syntax") 2196 2197 return f"ALTER COLUMN {this} DROP DEFAULT" 2198 2199 def renametable_sql(self, expression: exp.RenameTable) -> str: 2200 if not self.RENAME_TABLE_WITH_DB: 2201 # Remove db from tables 2202 expression = expression.transform( 2203 lambda n: exp.table_(n.this) if isinstance(n, exp.Table) else n 2204 ) 2205 this = self.sql(expression, "this") 2206 return f"RENAME TO {this}" 2207 2208 def altertable_sql(self, expression: exp.AlterTable) -> str: 2209 actions = expression.args["actions"] 2210 2211 if isinstance(actions[0], exp.ColumnDef): 2212 actions = self.expressions(expression, key="actions", prefix="ADD COLUMN ") 2213 elif isinstance(actions[0], exp.Schema): 2214 actions = self.expressions(expression, key="actions", prefix="ADD COLUMNS ") 2215 elif isinstance(actions[0], exp.Delete): 2216 actions = self.expressions(expression, key="actions", flat=True) 2217 else: 2218 actions = self.expressions(expression, key="actions") 2219 2220 exists = " IF EXISTS" if expression.args.get("exists") else "" 2221 return f"ALTER TABLE{exists} {self.sql(expression, 'this')} {actions}" 2222 2223 def droppartition_sql(self, expression: exp.DropPartition) -> str: 2224 expressions = self.expressions(expression) 2225 exists = " IF EXISTS " if expression.args.get("exists") else " " 2226 return f"DROP{exists}{expressions}" 2227 2228 def addconstraint_sql(self, expression: exp.AddConstraint) -> str: 2229 this = self.sql(expression, "this") 2230 expression_ = self.sql(expression, "expression") 2231 add_constraint = f"ADD CONSTRAINT {this}" if this else "ADD" 2232 2233 enforced = expression.args.get("enforced") 2234 if enforced is not None: 2235 return f"{add_constraint} CHECK ({expression_}){' ENFORCED' if enforced else ''}" 2236 2237 return f"{add_constraint} {expression_}" 2238 2239 def distinct_sql(self, expression: exp.Distinct) -> str: 2240 this = self.expressions(expression, flat=True) 2241 this = f" {this}" if this else "" 2242 2243 on = self.sql(expression, "on") 2244 on = f" ON {on}" if on else "" 2245 return f"DISTINCT{this}{on}" 2246 2247 def ignorenulls_sql(self, expression: exp.IgnoreNulls) -> str: 2248 return f"{self.sql(expression, 'this')} IGNORE NULLS" 2249 2250 def respectnulls_sql(self, expression: exp.RespectNulls) -> str: 2251 return f"{self.sql(expression, 'this')} RESPECT NULLS" 2252 2253 def intdiv_sql(self, expression: exp.IntDiv) -> str: 2254 return self.sql( 2255 exp.Cast( 2256 this=exp.Div(this=expression.this, expression=expression.expression), 2257 to=exp.DataType(this=exp.DataType.Type.INT), 2258 ) 2259 ) 2260 2261 def dpipe_sql(self, expression: exp.DPipe) -> str: 2262 return self.binary(expression, "||") 2263 2264 def safedpipe_sql(self, expression: exp.SafeDPipe) -> str: 2265 if self.STRICT_STRING_CONCAT: 2266 return self.func("CONCAT", *(exp.cast(e, "text") for e in expression.flatten())) 2267 return self.dpipe_sql(expression) 2268 2269 def div_sql(self, expression: exp.Div) -> str: 2270 return self.binary(expression, "/") 2271 2272 def overlaps_sql(self, expression: exp.Overlaps) -> str: 2273 return self.binary(expression, "OVERLAPS") 2274 2275 def distance_sql(self, expression: exp.Distance) -> str: 2276 return self.binary(expression, "<->") 2277 2278 def dot_sql(self, expression: exp.Dot) -> str: 2279 return f"{self.sql(expression, 'this')}.{self.sql(expression, 'expression')}" 2280 2281 def eq_sql(self, expression: exp.EQ) -> str: 2282 return self.binary(expression, "=") 2283 2284 def escape_sql(self, expression: exp.Escape) -> str: 2285 return self.binary(expression, "ESCAPE") 2286 2287 def glob_sql(self, expression: exp.Glob) -> str: 2288 return self.binary(expression, "GLOB") 2289 2290 def gt_sql(self, expression: exp.GT) -> str: 2291 return self.binary(expression, ">") 2292 2293 def gte_sql(self, expression: exp.GTE) -> str: 2294 return self.binary(expression, ">=") 2295 2296 def ilike_sql(self, expression: exp.ILike) -> str: 2297 return self.binary(expression, "ILIKE") 2298 2299 def ilikeany_sql(self, expression: exp.ILikeAny) -> str: 2300 return self.binary(expression, "ILIKE ANY") 2301 2302 def is_sql(self, expression: exp.Is) -> str: 2303 if not self.IS_BOOL_ALLOWED and isinstance(expression.expression, exp.Boolean): 2304 return self.sql( 2305 expression.this if expression.expression.this else exp.not_(expression.this) 2306 ) 2307 return self.binary(expression, "IS") 2308 2309 def like_sql(self, expression: exp.Like) -> str: 2310 return self.binary(expression, "LIKE") 2311 2312 def likeany_sql(self, expression: exp.LikeAny) -> str: 2313 return self.binary(expression, "LIKE ANY") 2314 2315 def similarto_sql(self, expression: exp.SimilarTo) -> str: 2316 return self.binary(expression, "SIMILAR TO") 2317 2318 def lt_sql(self, expression: exp.LT) -> str: 2319 return self.binary(expression, "<") 2320 2321 def lte_sql(self, expression: exp.LTE) -> str: 2322 return self.binary(expression, "<=") 2323 2324 def mod_sql(self, expression: exp.Mod) -> str: 2325 return self.binary(expression, "%") 2326 2327 def mul_sql(self, expression: exp.Mul) -> str: 2328 return self.binary(expression, "*") 2329 2330 def neq_sql(self, expression: exp.NEQ) -> str: 2331 return self.binary(expression, "<>") 2332 2333 def nullsafeeq_sql(self, expression: exp.NullSafeEQ) -> str: 2334 return self.binary(expression, "IS NOT DISTINCT FROM") 2335 2336 def nullsafeneq_sql(self, expression: exp.NullSafeNEQ) -> str: 2337 return self.binary(expression, "IS DISTINCT FROM") 2338 2339 def or_sql(self, expression: exp.Or) -> str: 2340 return self.connector_sql(expression, "OR") 2341 2342 def slice_sql(self, expression: exp.Slice) -> str: 2343 return self.binary(expression, ":") 2344 2345 def sub_sql(self, expression: exp.Sub) -> str: 2346 return self.binary(expression, "-") 2347 2348 def trycast_sql(self, expression: exp.TryCast) -> str: 2349 return self.cast_sql(expression, safe_prefix="TRY_") 2350 2351 def use_sql(self, expression: exp.Use) -> str: 2352 kind = self.sql(expression, "kind") 2353 kind = f" {kind}" if kind else "" 2354 this = self.sql(expression, "this") 2355 this = f" {this}" if this else "" 2356 return f"USE{kind}{this}" 2357 2358 def binary(self, expression: exp.Binary, op: str) -> str: 2359 op = self.maybe_comment(op, comments=expression.comments) 2360 return f"{self.sql(expression, 'this')} {op} {self.sql(expression, 'expression')}" 2361 2362 def function_fallback_sql(self, expression: exp.Func) -> str: 2363 args = [] 2364 2365 for key in expression.arg_types: 2366 arg_value = expression.args.get(key) 2367 2368 if isinstance(arg_value, list): 2369 for value in arg_value: 2370 args.append(value) 2371 elif arg_value is not None: 2372 args.append(arg_value) 2373 2374 return self.func(expression.sql_name(), *args) 2375 2376 def func( 2377 self, 2378 name: str, 2379 *args: t.Optional[exp.Expression | str], 2380 prefix: str = "(", 2381 suffix: str = ")", 2382 ) -> str: 2383 return f"{self.normalize_func(name)}{prefix}{self.format_args(*args)}{suffix}" 2384 2385 def format_args(self, *args: t.Optional[str | exp.Expression]) -> str: 2386 arg_sqls = tuple(self.sql(arg) for arg in args if arg is not None) 2387 if self.pretty and self.text_width(arg_sqls) > self.max_text_width: 2388 return self.indent("\n" + f",\n".join(arg_sqls) + "\n", skip_first=True, skip_last=True) 2389 return ", ".join(arg_sqls) 2390 2391 def text_width(self, args: t.Iterable) -> int: 2392 return sum(len(arg) for arg in args) 2393 2394 def format_time(self, expression: exp.Expression) -> t.Optional[str]: 2395 return format_time( 2396 self.sql(expression, "format"), self.INVERSE_TIME_MAPPING, self.INVERSE_TIME_TRIE 2397 ) 2398 2399 def expressions( 2400 self, 2401 expression: t.Optional[exp.Expression] = None, 2402 key: t.Optional[str] = None, 2403 sqls: t.Optional[t.List[str]] = None, 2404 flat: bool = False, 2405 indent: bool = True, 2406 skip_first: bool = False, 2407 sep: str = ", ", 2408 prefix: str = "", 2409 ) -> str: 2410 expressions = expression.args.get(key or "expressions") if expression else sqls 2411 2412 if not expressions: 2413 return "" 2414 2415 if flat: 2416 return sep.join(self.sql(e) for e in expressions) 2417 2418 num_sqls = len(expressions) 2419 2420 # These are calculated once in case we have the leading_comma / pretty option set, correspondingly 2421 pad = " " * self.pad 2422 stripped_sep = sep.strip() 2423 2424 result_sqls = [] 2425 for i, e in enumerate(expressions): 2426 sql = self.sql(e, comment=False) 2427 comments = self.maybe_comment("", e) if isinstance(e, exp.Expression) else "" 2428 2429 if self.pretty: 2430 if self.leading_comma: 2431 result_sqls.append(f"{sep if i > 0 else pad}{prefix}{sql}{comments}") 2432 else: 2433 result_sqls.append( 2434 f"{prefix}{sql}{stripped_sep if i + 1 < num_sqls else ''}{comments}" 2435 ) 2436 else: 2437 result_sqls.append(f"{prefix}{sql}{comments}{sep if i + 1 < num_sqls else ''}") 2438 2439 result_sql = "\n".join(result_sqls) if self.pretty else "".join(result_sqls) 2440 return self.indent(result_sql, skip_first=skip_first) if indent else result_sql 2441 2442 def op_expressions(self, op: str, expression: exp.Expression, flat: bool = False) -> str: 2443 flat = flat or isinstance(expression.parent, exp.Properties) 2444 expressions_sql = self.expressions(expression, flat=flat) 2445 if flat: 2446 return f"{op} {expressions_sql}" 2447 return f"{self.seg(op)}{self.sep() if expressions_sql else ''}{expressions_sql}" 2448 2449 def naked_property(self, expression: exp.Property) -> str: 2450 property_name = exp.Properties.PROPERTY_TO_NAME.get(expression.__class__) 2451 if not property_name: 2452 self.unsupported(f"Unsupported property {expression.__class__.__name__}") 2453 return f"{property_name} {self.sql(expression, 'this')}" 2454 2455 def set_operation(self, expression: exp.Expression, op: str) -> str: 2456 this = self.sql(expression, "this") 2457 op = self.seg(op) 2458 return self.query_modifiers( 2459 expression, f"{this}{op}{self.sep()}{self.sql(expression, 'expression')}" 2460 ) 2461 2462 def tag_sql(self, expression: exp.Tag) -> str: 2463 return f"{expression.args.get('prefix')}{self.sql(expression.this)}{expression.args.get('postfix')}" 2464 2465 def token_sql(self, token_type: TokenType) -> str: 2466 return self.TOKEN_MAPPING.get(token_type, token_type.name) 2467 2468 def userdefinedfunction_sql(self, expression: exp.UserDefinedFunction) -> str: 2469 this = self.sql(expression, "this") 2470 expressions = self.no_identify(self.expressions, expression) 2471 expressions = ( 2472 self.wrap(expressions) if expression.args.get("wrapped") else f" {expressions}" 2473 ) 2474 return f"{this}{expressions}" 2475 2476 def joinhint_sql(self, expression: exp.JoinHint) -> str: 2477 this = self.sql(expression, "this") 2478 expressions = self.expressions(expression, flat=True) 2479 return f"{this}({expressions})" 2480 2481 def kwarg_sql(self, expression: exp.Kwarg) -> str: 2482 return self.binary(expression, "=>") 2483 2484 def when_sql(self, expression: exp.When) -> str: 2485 matched = "MATCHED" if expression.args["matched"] else "NOT MATCHED" 2486 source = " BY SOURCE" if self.MATCHED_BY_SOURCE and expression.args.get("source") else "" 2487 condition = self.sql(expression, "condition") 2488 condition = f" AND {condition}" if condition else "" 2489 2490 then_expression = expression.args.get("then") 2491 if isinstance(then_expression, exp.Insert): 2492 then = f"INSERT {self.sql(then_expression, 'this')}" 2493 if "expression" in then_expression.args: 2494 then += f" VALUES {self.sql(then_expression, 'expression')}" 2495 elif isinstance(then_expression, exp.Update): 2496 if isinstance(then_expression.args.get("expressions"), exp.Star): 2497 then = f"UPDATE {self.sql(then_expression, 'expressions')}" 2498 else: 2499 then = f"UPDATE SET {self.expressions(then_expression, flat=True)}" 2500 else: 2501 then = self.sql(then_expression) 2502 return f"WHEN {matched}{source}{condition} THEN {then}" 2503 2504 def merge_sql(self, expression: exp.Merge) -> str: 2505 this = self.sql(expression, "this") 2506 using = f"USING {self.sql(expression, 'using')}" 2507 on = f"ON {self.sql(expression, 'on')}" 2508 return f"MERGE INTO {this} {using} {on} {self.expressions(expression, sep=' ')}" 2509 2510 def tochar_sql(self, expression: exp.ToChar) -> str: 2511 if expression.args.get("format"): 2512 self.unsupported("Format argument unsupported for TO_CHAR/TO_VARCHAR function") 2513 2514 return self.sql(exp.cast(expression.this, "text")) 2515 2516 def dictproperty_sql(self, expression: exp.DictProperty) -> str: 2517 this = self.sql(expression, "this") 2518 kind = self.sql(expression, "kind") 2519 settings_sql = self.expressions(expression, key="settings", sep=" ") 2520 args = f"({self.sep('')}{settings_sql}{self.seg(')', sep='')}" if settings_sql else "()" 2521 return f"{this}({kind}{args})" 2522 2523 def dictrange_sql(self, expression: exp.DictRange) -> str: 2524 this = self.sql(expression, "this") 2525 max = self.sql(expression, "max") 2526 min = self.sql(expression, "min") 2527 return f"{this}(MIN {min} MAX {max})" 2528 2529 def dictsubproperty_sql(self, expression: exp.DictSubProperty) -> str: 2530 return f"{self.sql(expression, 'this')} {self.sql(expression, 'value')}" 2531 2532 def oncluster_sql(self, expression: exp.OnCluster) -> str: 2533 return "" 2534 2535 def clusteredbyproperty_sql(self, expression: exp.ClusteredByProperty) -> str: 2536 expressions = self.expressions(expression, key="expressions", flat=True) 2537 sorted_by = self.expressions(expression, key="sorted_by", flat=True) 2538 sorted_by = f" SORTED BY ({sorted_by})" if sorted_by else "" 2539 buckets = self.sql(expression, "buckets") 2540 return f"CLUSTERED BY ({expressions}){sorted_by} INTO {buckets} BUCKETS" 2541 2542 def anyvalue_sql(self, expression: exp.AnyValue) -> str: 2543 this = self.sql(expression, "this") 2544 having = self.sql(expression, "having") 2545 2546 if having: 2547 this = f"{this} HAVING {'MAX' if expression.args.get('max') else 'MIN'} {having}" 2548 2549 return self.func("ANY_VALUE", this) 2550 2551 def querytransform_sql(self, expression: exp.QueryTransform) -> str: 2552 transform = self.func("TRANSFORM", *expression.expressions) 2553 row_format_before = self.sql(expression, "row_format_before") 2554 row_format_before = f" {row_format_before}" if row_format_before else "" 2555 record_writer = self.sql(expression, "record_writer") 2556 record_writer = f" RECORDWRITER {record_writer}" if record_writer else "" 2557 using = f" USING {self.sql(expression, 'command_script')}" 2558 schema = self.sql(expression, "schema") 2559 schema = f" AS {schema}" if schema else "" 2560 row_format_after = self.sql(expression, "row_format_after") 2561 row_format_after = f" {row_format_after}" if row_format_after else "" 2562 record_reader = self.sql(expression, "record_reader") 2563 record_reader = f" RECORDREADER {record_reader}" if record_reader else "" 2564 return f"{transform}{row_format_before}{record_writer}{using}{schema}{row_format_after}{record_reader}"
Generator converts a given syntax tree to the corresponding SQL string.
Arguments:
- pretty: Whether or not to format the produced SQL string. Default: False.
- identify: Determines when an identifier should be quoted. Possible values are: False (default): Never quote, except in cases where it's mandatory by the dialect. True or 'always': Always quote. 'safe': Only quote identifiers that are case insensitive.
- normalize: Whether or not to normalize identifiers to lowercase. Default: False.
- pad: Determines the pad size in a formatted string. Default: 2.
- indent: Determines the indentation size in a formatted string. Default: 2.
- normalize_functions: Whether or not to normalize all function names. Possible values are: "upper" or True (default): Convert names to uppercase. "lower": Convert names to lowercase. False: Disables function name normalization.
- unsupported_level: Determines the generator's behavior when it encounters unsupported expressions. Default ErrorLevel.WARN.
- max_unsupported: Maximum number of unsupported messages to include in a raised UnsupportedError. This is only relevant if unsupported_level is ErrorLevel.RAISE. Default: 3
- leading_comma: Determines whether or not the comma is leading or trailing in select expressions. This is only relevant when generating in pretty mode. Default: False
- max_text_width: The max number of characters in a segment before creating new lines in pretty mode. The default is on the smaller end because the length only represents a segment and not the true line length. Default: 80
- comments: Whether or not to preserve comments in the output SQL code. Default: True
Generator( pretty: Optional[bool] = None, identify: str | bool = False, normalize: bool = False, pad: int = 2, indent: int = 2, normalize_functions: Union[str, bool, NoneType] = None, unsupported_level: sqlglot.errors.ErrorLevel = <ErrorLevel.WARN: 'WARN'>, max_unsupported: int = 3, leading_comma: bool = False, max_text_width: int = 80, comments: bool = True)
334 def __init__( 335 self, 336 pretty: t.Optional[bool] = None, 337 identify: str | bool = False, 338 normalize: bool = False, 339 pad: int = 2, 340 indent: int = 2, 341 normalize_functions: t.Optional[str | bool] = None, 342 unsupported_level: ErrorLevel = ErrorLevel.WARN, 343 max_unsupported: int = 3, 344 leading_comma: bool = False, 345 max_text_width: int = 80, 346 comments: bool = True, 347 ): 348 import sqlglot 349 350 self.pretty = pretty if pretty is not None else sqlglot.pretty 351 self.identify = identify 352 self.normalize = normalize 353 self.pad = pad 354 self._indent = indent 355 self.unsupported_level = unsupported_level 356 self.max_unsupported = max_unsupported 357 self.leading_comma = leading_comma 358 self.max_text_width = max_text_width 359 self.comments = comments 360 361 # This is both a Dialect property and a Generator argument, so we prioritize the latter 362 self.normalize_functions = ( 363 self.NORMALIZE_FUNCTIONS if normalize_functions is None else normalize_functions 364 ) 365 366 self.unsupported_messages: t.List[str] = [] 367 self._escaped_quote_end: str = self.STRING_ESCAPE + self.QUOTE_END 368 self._escaped_identifier_end: str = self.IDENTIFIER_ESCAPE + self.IDENTIFIER_END 369 self._cache: t.Optional[t.Dict[int, str]] = None
TRANSFORMS =
{<class 'sqlglot.expressions.DateAdd'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.TsOrDsAdd'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CaseSpecificColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CharacterSetColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CharacterSetProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CheckColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CollateColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CopyGrantsProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CommentColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.DateFormatColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.DefaultColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.EncodeColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ExecuteAsProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ExternalProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.InlineLengthColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.LanguageProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.LocationProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.LogProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.MaterializedProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.NoPrimaryIndexProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.OnCommitProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.OnUpdateColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.PathColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ReturnsProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SetProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SettingsProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SqlSecurityProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.StabilityProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.TemporaryProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ToTableProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.TransientProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.TitleColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.UppercaseColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.VarMap'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.VolatileProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.WithJournalTableProperty'>: <function Generator.<lambda>>}
TYPE_MAPPING =
{<Type.NCHAR: 'NCHAR'>: 'CHAR', <Type.NVARCHAR: 'NVARCHAR'>: 'VARCHAR', <Type.MEDIUMTEXT: 'MEDIUMTEXT'>: 'TEXT', <Type.LONGTEXT: 'LONGTEXT'>: 'TEXT', <Type.MEDIUMBLOB: 'MEDIUMBLOB'>: 'BLOB', <Type.LONGBLOB: 'LONGBLOB'>: 'BLOB', <Type.INET: 'INET'>: 'INET'}
TIME_PART_SINGULARS =
{'microseconds': 'microsecond', 'seconds': 'second', 'minutes': 'minute', 'hours': 'hour', 'days': 'day', 'weeks': 'week', 'months': 'month', 'quarters': 'quarter', 'years': 'year'}
PROPERTIES_LOCATION =
{<class 'sqlglot.expressions.AlgorithmProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.AutoIncrementProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.BlockCompressionProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.CharacterSetProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.ChecksumProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.CollateProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.CopyGrantsProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.Cluster'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.ClusteredByProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.DataBlocksizeProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.DefinerProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.DictRange'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.DictProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.DistKeyProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.DistStyleProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.EngineProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.ExecuteAsProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.ExternalProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.FallbackProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.FileFormatProperty'>: <Location.POST_WITH: 'POST_WITH'>, <class 'sqlglot.expressions.FreespaceProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.IsolatedLoadingProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.JournalProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.LanguageProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.LikeProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.LocationProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.LockingProperty'>: <Location.POST_ALIAS: 'POST_ALIAS'>, <class 'sqlglot.expressions.LogProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.MaterializedProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.MergeBlockRatioProperty'>: <Location.POST_NAME: 'POST_NAME'>, <class 'sqlglot.expressions.NoPrimaryIndexProperty'>: <Location.POST_EXPRESSION: 'POST_EXPRESSION'>, <class 'sqlglot.expressions.OnCommitProperty'>: <Location.POST_EXPRESSION: 'POST_EXPRESSION'>, <class 'sqlglot.expressions.Order'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.PartitionedByProperty'>: <Location.POST_WITH: 'POST_WITH'>, <class 'sqlglot.expressions.PrimaryKey'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.Property'>: <Location.POST_WITH: 'POST_WITH'>, <class 'sqlglot.expressions.ReturnsProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.RowFormatProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.RowFormatDelimitedProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.RowFormatSerdeProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.SchemaCommentProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.SerdeProperties'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.Set'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.SettingsProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.SetProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.SortKeyProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.SqlSecurityProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.StabilityProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.TemporaryProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.ToTableProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.TransientProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.MergeTreeTTL'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.VolatileProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.WithDataProperty'>: <Location.POST_EXPRESSION: 'POST_EXPRESSION'>, <class 'sqlglot.expressions.WithJournalTableProperty'>: <Location.POST_NAME: 'POST_NAME'>}
WITH_SEPARATED_COMMENTS: Tuple[Type[sqlglot.expressions.Expression], ...] =
(<class 'sqlglot.expressions.Drop'>, <class 'sqlglot.expressions.From'>, <class 'sqlglot.expressions.Select'>, <class 'sqlglot.expressions.Where'>, <class 'sqlglot.expressions.With'>)
UNWRAPPED_INTERVAL_VALUES: Tuple[Type[sqlglot.expressions.Expression], ...] =
(<class 'sqlglot.expressions.Column'>, <class 'sqlglot.expressions.Literal'>, <class 'sqlglot.expressions.Neg'>, <class 'sqlglot.expressions.Paren'>)
@classmethod
def
can_identify(text: str, identify: str | bool = 'safe') -> bool:
248 @classmethod 249 def can_identify(cls, text: str, identify: str | bool = "safe") -> bool: 250 """Checks if text can be identified given an identify option. 251 252 Args: 253 text: The text to check. 254 identify: 255 "always" or `True`: Always returns true. 256 "safe": True if the identifier is case-insensitive. 257 258 Returns: 259 Whether or not the given text can be identified. 260 """ 261 if identify is True or identify == "always": 262 return True 263 264 if identify == "safe": 265 return not cls.case_sensitive(text) 266 267 return False
Checks if text can be identified given an identify option.
Arguments:
- text: The text to check.
- identify: "always" or
True
: Always returns true. "safe": True if the identifier is case-insensitive.
Returns:
Whether or not the given text can be identified.
def
generate( self, expression: Optional[sqlglot.expressions.Expression], cache: Optional[Dict[int, str]] = None) -> str:
371 def generate( 372 self, 373 expression: t.Optional[exp.Expression], 374 cache: t.Optional[t.Dict[int, str]] = None, 375 ) -> str: 376 """ 377 Generates the SQL string corresponding to the given syntax tree. 378 379 Args: 380 expression: The syntax tree. 381 cache: An optional sql string cache. This leverages the hash of an Expression 382 which can be slow to compute, so only use it if you set _hash on each node. 383 384 Returns: 385 The SQL string corresponding to `expression`. 386 """ 387 if cache is not None: 388 self._cache = cache 389 390 self.unsupported_messages = [] 391 sql = self.sql(expression).strip() 392 self._cache = None 393 394 if self.unsupported_level == ErrorLevel.IGNORE: 395 return sql 396 397 if self.unsupported_level == ErrorLevel.WARN: 398 for msg in self.unsupported_messages: 399 logger.warning(msg) 400 elif self.unsupported_level == ErrorLevel.RAISE and self.unsupported_messages: 401 raise UnsupportedError(concat_messages(self.unsupported_messages, self.max_unsupported)) 402 403 if self.pretty: 404 sql = sql.replace(self.SENTINEL_LINE_BREAK, "\n") 405 return sql
Generates the SQL string corresponding to the given syntax tree.
Arguments:
- expression: The syntax tree.
- cache: An optional sql string cache. This leverages the hash of an Expression which can be slow to compute, so only use it if you set _hash on each node.
Returns:
The SQL string corresponding to
expression
.
def
maybe_comment( self, sql: str, expression: Optional[sqlglot.expressions.Expression] = None, comments: Optional[List[str]] = None) -> str:
423 def maybe_comment( 424 self, 425 sql: str, 426 expression: t.Optional[exp.Expression] = None, 427 comments: t.Optional[t.List[str]] = None, 428 ) -> str: 429 comments = ( 430 ((expression and expression.comments) if comments is None else comments) # type: ignore 431 if self.comments 432 else None 433 ) 434 435 if not comments or isinstance(expression, exp.Binary): 436 return sql 437 438 sep = "\n" if self.pretty else " " 439 comments_sql = sep.join( 440 f"/*{self.pad_comment(comment)}*/" for comment in comments if comment 441 ) 442 443 if not comments_sql: 444 return sql 445 446 if isinstance(expression, self.WITH_SEPARATED_COMMENTS): 447 return ( 448 f"{self.sep()}{comments_sql}{sql}" 449 if sql[0].isspace() 450 else f"{comments_sql}{self.sep()}{sql}" 451 ) 452 453 return f"{sql} {comments_sql}"
455 def wrap(self, expression: exp.Expression | str) -> str: 456 this_sql = self.indent( 457 self.sql(expression) 458 if isinstance(expression, (exp.Select, exp.Union)) 459 else self.sql(expression, "this"), 460 level=1, 461 pad=0, 462 ) 463 return f"({self.sep('')}{this_sql}{self.seg(')', sep='')}"
def
indent( self, sql: str, level: int = 0, pad: Optional[int] = None, skip_first: bool = False, skip_last: bool = False) -> str:
479 def indent( 480 self, 481 sql: str, 482 level: int = 0, 483 pad: t.Optional[int] = None, 484 skip_first: bool = False, 485 skip_last: bool = False, 486 ) -> str: 487 if not self.pretty: 488 return sql 489 490 pad = self.pad if pad is None else pad 491 lines = sql.split("\n") 492 493 return "\n".join( 494 line 495 if (skip_first and i == 0) or (skip_last and i == len(lines) - 1) 496 else f"{' ' * (level * self._indent + pad)}{line}" 497 for i, line in enumerate(lines) 498 )
def
sql( self, expression: Union[str, sqlglot.expressions.Expression, NoneType], key: Optional[str] = None, comment: bool = True) -> str:
500 def sql( 501 self, 502 expression: t.Optional[str | exp.Expression], 503 key: t.Optional[str] = None, 504 comment: bool = True, 505 ) -> str: 506 if not expression: 507 return "" 508 509 if isinstance(expression, str): 510 return expression 511 512 if key: 513 value = expression.args.get(key) 514 if value: 515 return self.sql(value) 516 return "" 517 518 if self._cache is not None: 519 expression_id = hash(expression) 520 521 if expression_id in self._cache: 522 return self._cache[expression_id] 523 524 transform = self.TRANSFORMS.get(expression.__class__) 525 526 if callable(transform): 527 sql = transform(self, expression) 528 elif transform: 529 sql = transform 530 elif isinstance(expression, exp.Expression): 531 exp_handler_name = f"{expression.key}_sql" 532 533 if hasattr(self, exp_handler_name): 534 sql = getattr(self, exp_handler_name)(expression) 535 elif isinstance(expression, exp.Func): 536 sql = self.function_fallback_sql(expression) 537 elif isinstance(expression, exp.Property): 538 sql = self.property_sql(expression) 539 else: 540 raise ValueError(f"Unsupported expression type {expression.__class__.__name__}") 541 else: 542 raise ValueError(f"Expected an Expression. Received {type(expression)}: {expression}") 543 544 sql = self.maybe_comment(sql, expression) if self.comments and comment else sql 545 546 if self._cache is not None: 547 self._cache[expression_id] = sql 548 return sql
555 def cache_sql(self, expression: exp.Cache) -> str: 556 lazy = " LAZY" if expression.args.get("lazy") else "" 557 table = self.sql(expression, "this") 558 options = expression.args.get("options") 559 options = f" OPTIONS({self.sql(options[0])} = {self.sql(options[1])})" if options else "" 560 sql = self.sql(expression, "expression") 561 sql = f" AS{self.sep()}{sql}" if sql else "" 562 sql = f"CACHE{lazy} TABLE {table}{options}{sql}" 563 return self.prepend_ctes(expression, sql)
565 def characterset_sql(self, expression: exp.CharacterSet) -> str: 566 if isinstance(expression.parent, exp.Cast): 567 return f"CHAR CHARACTER SET {self.sql(expression, 'this')}" 568 default = "DEFAULT " if expression.args.get("default") else "" 569 return f"{default}CHARACTER SET={self.sql(expression, 'this')}"
571 def column_sql(self, expression: exp.Column) -> str: 572 join_mark = " (+)" if expression.args.get("join_mark") else "" 573 574 if join_mark and not self.COLUMN_JOIN_MARKS_SUPPORTED: 575 join_mark = "" 576 self.unsupported("Outer join syntax using the (+) operator is not supported.") 577 578 column = ".".join( 579 self.sql(part) 580 for part in ( 581 expression.args.get("catalog"), 582 expression.args.get("db"), 583 expression.args.get("table"), 584 expression.args.get("this"), 585 ) 586 if part 587 ) 588 589 return f"{column}{join_mark}"
597 def columndef_sql(self, expression: exp.ColumnDef, sep: str = " ") -> str: 598 column = self.sql(expression, "this") 599 kind = self.sql(expression, "kind") 600 constraints = self.expressions(expression, key="constraints", sep=" ", flat=True) 601 exists = "IF NOT EXISTS " if expression.args.get("exists") else "" 602 kind = f"{sep}{kind}" if kind else "" 603 constraints = f" {constraints}" if constraints else "" 604 position = self.sql(expression, "position") 605 position = f" {position}" if position else "" 606 607 return f"{exists}{column}{kind}{constraints}{position}"
def
compresscolumnconstraint_sql(self, expression: sqlglot.expressions.CompressColumnConstraint) -> str:
def
generatedasidentitycolumnconstraint_sql( self, expression: sqlglot.expressions.GeneratedAsIdentityColumnConstraint) -> str:
625 def generatedasidentitycolumnconstraint_sql( 626 self, expression: exp.GeneratedAsIdentityColumnConstraint 627 ) -> str: 628 this = "" 629 if expression.this is not None: 630 on_null = "ON NULL " if expression.args.get("on_null") else "" 631 this = " ALWAYS " if expression.this else f" BY DEFAULT {on_null}" 632 633 start = expression.args.get("start") 634 start = f"START WITH {start}" if start else "" 635 increment = expression.args.get("increment") 636 increment = f" INCREMENT BY {increment}" if increment else "" 637 minvalue = expression.args.get("minvalue") 638 minvalue = f" MINVALUE {minvalue}" if minvalue else "" 639 maxvalue = expression.args.get("maxvalue") 640 maxvalue = f" MAXVALUE {maxvalue}" if maxvalue else "" 641 cycle = expression.args.get("cycle") 642 cycle_sql = "" 643 644 if cycle is not None: 645 cycle_sql = f"{' NO' if not cycle else ''} CYCLE" 646 cycle_sql = cycle_sql.strip() if not start and not increment else cycle_sql 647 648 sequence_opts = "" 649 if start or increment or cycle_sql: 650 sequence_opts = f"{start}{increment}{minvalue}{maxvalue}{cycle_sql}" 651 sequence_opts = f" ({sequence_opts.strip()})" 652 653 expr = self.sql(expression, "expression") 654 expr = f"({expr})" if expr else "IDENTITY" 655 656 return f"GENERATED{this}AS {expr}{sequence_opts}"
def
notnullcolumnconstraint_sql(self, expression: sqlglot.expressions.NotNullColumnConstraint) -> str:
def
primarykeycolumnconstraint_sql(self, expression: sqlglot.expressions.PrimaryKeyColumnConstraint) -> str:
def
uniquecolumnconstraint_sql(self, expression: sqlglot.expressions.UniqueColumnConstraint) -> str:
def
createable_sql( self, expression: sqlglot.expressions.Create, locations: dict[sqlglot.expressions.Properties.Location, list[sqlglot.expressions.Property]]) -> str:
677 def create_sql(self, expression: exp.Create) -> str: 678 kind = self.sql(expression, "kind").upper() 679 properties = expression.args.get("properties") 680 properties_locs = self.locate_properties(properties) if properties else {} 681 682 this = self.createable_sql(expression, properties_locs) 683 684 properties_sql = "" 685 if properties_locs.get(exp.Properties.Location.POST_SCHEMA) or properties_locs.get( 686 exp.Properties.Location.POST_WITH 687 ): 688 properties_sql = self.sql( 689 exp.Properties( 690 expressions=[ 691 *properties_locs[exp.Properties.Location.POST_SCHEMA], 692 *properties_locs[exp.Properties.Location.POST_WITH], 693 ] 694 ) 695 ) 696 697 begin = " BEGIN" if expression.args.get("begin") else "" 698 expression_sql = self.sql(expression, "expression") 699 if expression_sql: 700 expression_sql = f"{begin}{self.sep()}{expression_sql}" 701 702 if self.CREATE_FUNCTION_RETURN_AS or not isinstance(expression.expression, exp.Return): 703 if properties_locs.get(exp.Properties.Location.POST_ALIAS): 704 postalias_props_sql = self.properties( 705 exp.Properties( 706 expressions=properties_locs[exp.Properties.Location.POST_ALIAS] 707 ), 708 wrapped=False, 709 ) 710 expression_sql = f" AS {postalias_props_sql}{expression_sql}" 711 else: 712 expression_sql = f" AS{expression_sql}" 713 714 postindex_props_sql = "" 715 if properties_locs.get(exp.Properties.Location.POST_INDEX): 716 postindex_props_sql = self.properties( 717 exp.Properties(expressions=properties_locs[exp.Properties.Location.POST_INDEX]), 718 wrapped=False, 719 prefix=" ", 720 ) 721 722 indexes = self.expressions(expression, key="indexes", indent=False, sep=" ") 723 indexes = f" {indexes}" if indexes else "" 724 index_sql = indexes + postindex_props_sql 725 726 replace = " OR REPLACE" if expression.args.get("replace") else "" 727 unique = " UNIQUE" if expression.args.get("unique") else "" 728 729 postcreate_props_sql = "" 730 if properties_locs.get(exp.Properties.Location.POST_CREATE): 731 postcreate_props_sql = self.properties( 732 exp.Properties(expressions=properties_locs[exp.Properties.Location.POST_CREATE]), 733 sep=" ", 734 prefix=" ", 735 wrapped=False, 736 ) 737 738 modifiers = "".join((replace, unique, postcreate_props_sql)) 739 740 postexpression_props_sql = "" 741 if properties_locs.get(exp.Properties.Location.POST_EXPRESSION): 742 postexpression_props_sql = self.properties( 743 exp.Properties( 744 expressions=properties_locs[exp.Properties.Location.POST_EXPRESSION] 745 ), 746 sep=" ", 747 prefix=" ", 748 wrapped=False, 749 ) 750 751 exists_sql = " IF NOT EXISTS" if expression.args.get("exists") else "" 752 no_schema_binding = ( 753 " WITH NO SCHEMA BINDING" if expression.args.get("no_schema_binding") else "" 754 ) 755 756 clone = self.sql(expression, "clone") 757 clone = f" {clone}" if clone else "" 758 759 expression_sql = f"CREATE{modifiers} {kind}{exists_sql} {this}{properties_sql}{expression_sql}{postexpression_props_sql}{index_sql}{no_schema_binding}{clone}" 760 return self.prepend_ctes(expression, expression_sql)
762 def clone_sql(self, expression: exp.Clone) -> str: 763 this = self.sql(expression, "this") 764 when = self.sql(expression, "when") 765 766 if when: 767 kind = self.sql(expression, "kind") 768 expr = self.sql(expression, "expression") 769 return f"CLONE {this} {when} ({kind} => {expr})" 770 771 return f"CLONE {this}"
826 def datatype_sql(self, expression: exp.DataType) -> str: 827 type_value = expression.this 828 type_sql = ( 829 self.TYPE_MAPPING.get(type_value, type_value.value) 830 if isinstance(type_value, exp.DataType.Type) 831 else type_value 832 ) 833 nested = "" 834 interior = self.expressions(expression, flat=True) 835 values = "" 836 if interior: 837 if expression.args.get("nested"): 838 nested = f"{self.STRUCT_DELIMITER[0]}{interior}{self.STRUCT_DELIMITER[1]}" 839 if expression.args.get("values") is not None: 840 delimiters = ("[", "]") if type_value == exp.DataType.Type.ARRAY else ("(", ")") 841 values = self.expressions(expression, key="values", flat=True) 842 values = f"{delimiters[0]}{values}{delimiters[1]}" 843 else: 844 nested = f"({interior})" 845 846 return f"{type_sql}{nested}{values}"
848 def directory_sql(self, expression: exp.Directory) -> str: 849 local = "LOCAL " if expression.args.get("local") else "" 850 row_format = self.sql(expression, "row_format") 851 row_format = f" {row_format}" if row_format else "" 852 return f"{local}DIRECTORY {self.sql(expression, 'this')}{row_format}"
854 def delete_sql(self, expression: exp.Delete) -> str: 855 this = self.sql(expression, "this") 856 this = f" FROM {this}" if this else "" 857 using = self.sql(expression, "using") 858 using = f" USING {using}" if using else "" 859 where = self.sql(expression, "where") 860 returning = self.sql(expression, "returning") 861 limit = self.sql(expression, "limit") 862 tables = self.expressions(expression, key="tables") 863 tables = f" {tables}" if tables else "" 864 if self.RETURNING_END: 865 expression_sql = f"{this}{using}{where}{returning}{limit}" 866 else: 867 expression_sql = f"{returning}{this}{using}{where}{limit}" 868 return self.prepend_ctes(expression, f"DELETE{tables}{expression_sql}")
870 def drop_sql(self, expression: exp.Drop) -> str: 871 this = self.sql(expression, "this") 872 kind = expression.args["kind"] 873 exists_sql = " IF EXISTS " if expression.args.get("exists") else " " 874 temporary = " TEMPORARY" if expression.args.get("temporary") else "" 875 materialized = " MATERIALIZED" if expression.args.get("materialized") else "" 876 cascade = " CASCADE" if expression.args.get("cascade") else "" 877 constraints = " CONSTRAINTS" if expression.args.get("constraints") else "" 878 purge = " PURGE" if expression.args.get("purge") else "" 879 return ( 880 f"DROP{temporary}{materialized} {kind}{exists_sql}{this}{cascade}{constraints}{purge}" 881 )
892 def fetch_sql(self, expression: exp.Fetch) -> str: 893 direction = expression.args.get("direction") 894 direction = f" {direction.upper()}" if direction else "" 895 count = expression.args.get("count") 896 count = f" {count}" if count else "" 897 if expression.args.get("percent"): 898 count = f"{count} PERCENT" 899 with_ties_or_only = "WITH TIES" if expression.args.get("with_ties") else "ONLY" 900 return f"{self.seg('FETCH')}{direction}{count} ROWS {with_ties_or_only}"
914 def index_sql(self, expression: exp.Index) -> str: 915 unique = "UNIQUE " if expression.args.get("unique") else "" 916 primary = "PRIMARY " if expression.args.get("primary") else "" 917 amp = "AMP " if expression.args.get("amp") else "" 918 name = self.sql(expression, "this") 919 name = f"{name} " if name else "" 920 table = self.sql(expression, "table") 921 table = f"{self.INDEX_ON} {table} " if table else "" 922 using = self.sql(expression, "using") 923 using = f"USING {using} " if using else "" 924 index = "INDEX " if not table else "" 925 columns = self.expressions(expression, key="columns", flat=True) 926 columns = f"({columns})" if columns else "" 927 partition_by = self.expressions(expression, key="partition_by", flat=True) 928 partition_by = f" PARTITION BY {partition_by}" if partition_by else "" 929 return f"{unique}{primary}{amp}{index}{name}{table}{using}{columns}{partition_by}"
931 def identifier_sql(self, expression: exp.Identifier) -> str: 932 text = expression.name 933 lower = text.lower() 934 text = lower if self.normalize and not expression.quoted else text 935 text = text.replace(self.IDENTIFIER_END, self._escaped_identifier_end) 936 if ( 937 expression.quoted 938 or self.can_identify(text, self.identify) 939 or lower in self.RESERVED_KEYWORDS 940 or (not self.IDENTIFIERS_CAN_START_WITH_DIGIT and text[:1].isdigit()) 941 ): 942 text = f"{self.IDENTIFIER_START}{text}{self.IDENTIFIER_END}" 943 return text
945 def inputoutputformat_sql(self, expression: exp.InputOutputFormat) -> str: 946 input_format = self.sql(expression, "input_format") 947 input_format = f"INPUTFORMAT {input_format}" if input_format else "" 948 output_format = self.sql(expression, "output_format") 949 output_format = f"OUTPUTFORMAT {output_format}" if output_format else "" 950 return self.sep().join((input_format, output_format))
959 def properties_sql(self, expression: exp.Properties) -> str: 960 root_properties = [] 961 with_properties = [] 962 963 for p in expression.expressions: 964 p_loc = self.PROPERTIES_LOCATION[p.__class__] 965 if p_loc == exp.Properties.Location.POST_WITH: 966 with_properties.append(p) 967 elif p_loc == exp.Properties.Location.POST_SCHEMA: 968 root_properties.append(p) 969 970 return self.root_properties( 971 exp.Properties(expressions=root_properties) 972 ) + self.with_properties(exp.Properties(expressions=with_properties))
def
properties( self, properties: sqlglot.expressions.Properties, prefix: str = '', sep: str = ', ', suffix: str = '', wrapped: bool = True) -> str:
979 def properties( 980 self, 981 properties: exp.Properties, 982 prefix: str = "", 983 sep: str = ", ", 984 suffix: str = "", 985 wrapped: bool = True, 986 ) -> str: 987 if properties.expressions: 988 expressions = self.expressions(properties, sep=sep, indent=False) 989 expressions = self.wrap(expressions) if wrapped else expressions 990 return f"{prefix}{' ' if prefix and prefix != ' ' else ''}{expressions}{suffix}" 991 return ""
def
locate_properties( self, properties: sqlglot.expressions.Properties) -> Dict[sqlglot.expressions.Properties.Location, list[sqlglot.expressions.Property]]:
996 def locate_properties( 997 self, properties: exp.Properties 998 ) -> t.Dict[exp.Properties.Location, list[exp.Property]]: 999 properties_locs: t.Dict[exp.Properties.Location, list[exp.Property]] = { 1000 key: [] for key in exp.Properties.Location 1001 } 1002 1003 for p in properties.expressions: 1004 p_loc = self.PROPERTIES_LOCATION[p.__class__] 1005 if p_loc == exp.Properties.Location.POST_NAME: 1006 properties_locs[exp.Properties.Location.POST_NAME].append(p) 1007 elif p_loc == exp.Properties.Location.POST_INDEX: 1008 properties_locs[exp.Properties.Location.POST_INDEX].append(p) 1009 elif p_loc == exp.Properties.Location.POST_SCHEMA: 1010 properties_locs[exp.Properties.Location.POST_SCHEMA].append(p) 1011 elif p_loc == exp.Properties.Location.POST_WITH: 1012 properties_locs[exp.Properties.Location.POST_WITH].append(p) 1013 elif p_loc == exp.Properties.Location.POST_CREATE: 1014 properties_locs[exp.Properties.Location.POST_CREATE].append(p) 1015 elif p_loc == exp.Properties.Location.POST_ALIAS: 1016 properties_locs[exp.Properties.Location.POST_ALIAS].append(p) 1017 elif p_loc == exp.Properties.Location.POST_EXPRESSION: 1018 properties_locs[exp.Properties.Location.POST_EXPRESSION].append(p) 1019 elif p_loc == exp.Properties.Location.UNSUPPORTED: 1020 self.unsupported(f"Unsupported property {p.key}") 1021 1022 return properties_locs
1024 def property_sql(self, expression: exp.Property) -> str: 1025 property_cls = expression.__class__ 1026 if property_cls == exp.Property: 1027 return f"{expression.name}={self.sql(expression, 'value')}" 1028 1029 property_name = exp.Properties.PROPERTY_TO_NAME.get(property_cls) 1030 if not property_name: 1031 self.unsupported(f"Unsupported property {expression.key}") 1032 1033 return f"{property_name}={self.sql(expression, 'this')}"
1045 def journalproperty_sql(self, expression: exp.JournalProperty) -> str: 1046 no = "NO " if expression.args.get("no") else "" 1047 local = expression.args.get("local") 1048 local = f"{local} " if local else "" 1049 dual = "DUAL " if expression.args.get("dual") else "" 1050 before = "BEFORE " if expression.args.get("before") else "" 1051 after = "AFTER " if expression.args.get("after") else "" 1052 return f"{no}{local}{dual}{before}{after}JOURNAL"
def
mergeblockratioproperty_sql(self, expression: sqlglot.expressions.MergeBlockRatioProperty) -> str:
1068 def mergeblockratioproperty_sql(self, expression: exp.MergeBlockRatioProperty) -> str: 1069 if expression.args.get("no"): 1070 return "NO MERGEBLOCKRATIO" 1071 if expression.args.get("default"): 1072 return "DEFAULT MERGEBLOCKRATIO" 1073 1074 percent = " PERCENT" if expression.args.get("percent") else "" 1075 return f"MERGEBLOCKRATIO={self.sql(expression, 'this')}{percent}"
1077 def datablocksizeproperty_sql(self, expression: exp.DataBlocksizeProperty) -> str: 1078 default = expression.args.get("default") 1079 minimum = expression.args.get("minimum") 1080 maximum = expression.args.get("maximum") 1081 if default or minimum or maximum: 1082 if default: 1083 prop = "DEFAULT" 1084 elif minimum: 1085 prop = "MINIMUM" 1086 else: 1087 prop = "MAXIMUM" 1088 return f"{prop} DATABLOCKSIZE" 1089 units = expression.args.get("units") 1090 units = f" {units}" if units else "" 1091 return f"DATABLOCKSIZE={self.sql(expression, 'size')}{units}"
def
blockcompressionproperty_sql(self, expression: sqlglot.expressions.BlockCompressionProperty) -> str:
1093 def blockcompressionproperty_sql(self, expression: exp.BlockCompressionProperty) -> str: 1094 autotemp = expression.args.get("autotemp") 1095 always = expression.args.get("always") 1096 default = expression.args.get("default") 1097 manual = expression.args.get("manual") 1098 never = expression.args.get("never") 1099 1100 if autotemp is not None: 1101 prop = f"AUTOTEMP({self.expressions(autotemp)})" 1102 elif always: 1103 prop = "ALWAYS" 1104 elif default: 1105 prop = "DEFAULT" 1106 elif manual: 1107 prop = "MANUAL" 1108 elif never: 1109 prop = "NEVER" 1110 return f"BLOCKCOMPRESSION={prop}"
def
isolatedloadingproperty_sql(self, expression: sqlglot.expressions.IsolatedLoadingProperty) -> str:
1112 def isolatedloadingproperty_sql(self, expression: exp.IsolatedLoadingProperty) -> str: 1113 no = expression.args.get("no") 1114 no = " NO" if no else "" 1115 concurrent = expression.args.get("concurrent") 1116 concurrent = " CONCURRENT" if concurrent else "" 1117 1118 for_ = "" 1119 if expression.args.get("for_all"): 1120 for_ = " FOR ALL" 1121 elif expression.args.get("for_insert"): 1122 for_ = " FOR INSERT" 1123 elif expression.args.get("for_none"): 1124 for_ = " FOR NONE" 1125 return f"WITH{no}{concurrent} ISOLATED LOADING{for_}"
1127 def lockingproperty_sql(self, expression: exp.LockingProperty) -> str: 1128 kind = expression.args.get("kind") 1129 this = f" {self.sql(expression, 'this')}" if expression.this else "" 1130 for_or_in = expression.args.get("for_or_in") 1131 lock_type = expression.args.get("lock_type") 1132 override = " OVERRIDE" if expression.args.get("override") else "" 1133 return f"LOCKING {kind}{this} {for_or_in} {lock_type}{override}"
1135 def withdataproperty_sql(self, expression: exp.WithDataProperty) -> str: 1136 data_sql = f"WITH {'NO ' if expression.args.get('no') else ''}DATA" 1137 statistics = expression.args.get("statistics") 1138 statistics_sql = "" 1139 if statistics is not None: 1140 statistics_sql = f" AND {'NO ' if not statistics else ''}STATISTICS" 1141 return f"{data_sql}{statistics_sql}"
1143 def insert_sql(self, expression: exp.Insert) -> str: 1144 overwrite = expression.args.get("overwrite") 1145 1146 if isinstance(expression.this, exp.Directory): 1147 this = " OVERWRITE" if overwrite else " INTO" 1148 else: 1149 this = " OVERWRITE TABLE" if overwrite else " INTO" 1150 1151 alternative = expression.args.get("alternative") 1152 alternative = f" OR {alternative}" if alternative else "" 1153 ignore = " IGNORE" if expression.args.get("ignore") else "" 1154 1155 this = f"{this} {self.sql(expression, 'this')}" 1156 1157 exists = " IF EXISTS" if expression.args.get("exists") else "" 1158 partition_sql = ( 1159 f" {self.sql(expression, 'partition')}" if expression.args.get("partition") else "" 1160 ) 1161 where = self.sql(expression, "where") 1162 where = f"{self.sep()}REPLACE WHERE {where}" if where else "" 1163 expression_sql = f"{self.sep()}{self.sql(expression, 'expression')}" 1164 conflict = self.sql(expression, "conflict") 1165 returning = self.sql(expression, "returning") 1166 1167 if self.RETURNING_END: 1168 expression_sql = f"{expression_sql}{conflict}{returning}" 1169 else: 1170 expression_sql = f"{returning}{expression_sql}{conflict}" 1171 1172 sql = f"INSERT{alternative}{ignore}{this}{exists}{partition_sql}{where}{expression_sql}" 1173 return self.prepend_ctes(expression, sql)
1190 def onconflict_sql(self, expression: exp.OnConflict) -> str: 1191 conflict = "ON DUPLICATE KEY" if expression.args.get("duplicate") else "ON CONFLICT" 1192 constraint = self.sql(expression, "constraint") 1193 if constraint: 1194 constraint = f"ON CONSTRAINT {constraint}" 1195 key = self.expressions(expression, key="key", flat=True) 1196 do = "" if expression.args.get("duplicate") else " DO " 1197 nothing = "NOTHING" if expression.args.get("nothing") else "" 1198 expressions = self.expressions(expression, flat=True) 1199 set_keyword = "SET " if self.DUPLICATE_KEY_UPDATE_WITH_SET else "" 1200 if expressions: 1201 expressions = f"UPDATE {set_keyword}{expressions}" 1202 return f"{self.seg(conflict)} {constraint}{key}{do}{nothing}{expressions}"
def
rowformatdelimitedproperty_sql(self, expression: sqlglot.expressions.RowFormatDelimitedProperty) -> str:
1207 def rowformatdelimitedproperty_sql(self, expression: exp.RowFormatDelimitedProperty) -> str: 1208 fields = expression.args.get("fields") 1209 fields = f" FIELDS TERMINATED BY {fields}" if fields else "" 1210 escaped = expression.args.get("escaped") 1211 escaped = f" ESCAPED BY {escaped}" if escaped else "" 1212 items = expression.args.get("collection_items") 1213 items = f" COLLECTION ITEMS TERMINATED BY {items}" if items else "" 1214 keys = expression.args.get("map_keys") 1215 keys = f" MAP KEYS TERMINATED BY {keys}" if keys else "" 1216 lines = expression.args.get("lines") 1217 lines = f" LINES TERMINATED BY {lines}" if lines else "" 1218 null = expression.args.get("null") 1219 null = f" NULL DEFINED AS {null}" if null else "" 1220 return f"ROW FORMAT DELIMITED{fields}{escaped}{items}{keys}{lines}{null}"
1231 def table_sql(self, expression: exp.Table, sep: str = " AS ") -> str: 1232 table = ".".join( 1233 part 1234 for part in [ 1235 self.sql(expression, "catalog"), 1236 self.sql(expression, "db"), 1237 self.sql(expression, "this"), 1238 ] 1239 if part 1240 ) 1241 1242 alias = self.sql(expression, "alias") 1243 alias = f"{sep}{alias}" if alias else "" 1244 hints = self.expressions(expression, key="hints", sep=" ") 1245 hints = f" {hints}" if hints and self.TABLE_HINTS else "" 1246 pivots = self.expressions(expression, key="pivots", sep=" ", flat=True) 1247 pivots = f" {pivots}" if pivots else "" 1248 joins = self.expressions(expression, key="joins", sep="", skip_first=True) 1249 laterals = self.expressions(expression, key="laterals", sep="") 1250 system_time = expression.args.get("system_time") 1251 system_time = f" {self.sql(expression, 'system_time')}" if system_time else "" 1252 1253 return f"{table}{system_time}{alias}{hints}{pivots}{joins}{laterals}"
def
tablesample_sql( self, expression: sqlglot.expressions.TableSample, seed_prefix: str = 'SEED', sep=' AS ') -> str:
1255 def tablesample_sql( 1256 self, expression: exp.TableSample, seed_prefix: str = "SEED", sep=" AS " 1257 ) -> str: 1258 if self.ALIAS_POST_TABLESAMPLE and expression.this.alias: 1259 table = expression.this.copy() 1260 table.set("alias", None) 1261 this = self.sql(table) 1262 alias = f"{sep}{self.sql(expression.this, 'alias')}" 1263 else: 1264 this = self.sql(expression, "this") 1265 alias = "" 1266 method = self.sql(expression, "method") 1267 method = f"{method.upper()} " if method and self.TABLESAMPLE_WITH_METHOD else "" 1268 numerator = self.sql(expression, "bucket_numerator") 1269 denominator = self.sql(expression, "bucket_denominator") 1270 field = self.sql(expression, "bucket_field") 1271 field = f" ON {field}" if field else "" 1272 bucket = f"BUCKET {numerator} OUT OF {denominator}{field}" if numerator else "" 1273 percent = self.sql(expression, "percent") 1274 percent = f"{percent} PERCENT" if percent else "" 1275 rows = self.sql(expression, "rows") 1276 rows = f"{rows} ROWS" if rows else "" 1277 size = self.sql(expression, "size") 1278 if size and self.TABLESAMPLE_SIZE_IS_PERCENT: 1279 size = f"{size} PERCENT" 1280 seed = self.sql(expression, "seed") 1281 seed = f" {seed_prefix} ({seed})" if seed else "" 1282 kind = expression.args.get("kind", "TABLESAMPLE") 1283 return f"{this} {kind} {method}({bucket}{percent}{rows}{size}){seed}{alias}"
1285 def pivot_sql(self, expression: exp.Pivot) -> str: 1286 expressions = self.expressions(expression, flat=True) 1287 1288 if expression.this: 1289 this = self.sql(expression, "this") 1290 on = f"{self.seg('ON')} {expressions}" 1291 using = self.expressions(expression, key="using", flat=True) 1292 using = f"{self.seg('USING')} {using}" if using else "" 1293 group = self.sql(expression, "group") 1294 return f"PIVOT {this}{on}{using}{group}" 1295 1296 alias = self.sql(expression, "alias") 1297 alias = f" AS {alias}" if alias else "" 1298 unpivot = expression.args.get("unpivot") 1299 direction = "UNPIVOT" if unpivot else "PIVOT" 1300 field = self.sql(expression, "field") 1301 return f"{direction}({expressions} FOR {field}){alias}"
1306 def update_sql(self, expression: exp.Update) -> str: 1307 this = self.sql(expression, "this") 1308 set_sql = self.expressions(expression, flat=True) 1309 from_sql = self.sql(expression, "from") 1310 where_sql = self.sql(expression, "where") 1311 returning = self.sql(expression, "returning") 1312 limit = self.sql(expression, "limit") 1313 if self.RETURNING_END: 1314 expression_sql = f"{from_sql}{where_sql}{returning}{limit}" 1315 else: 1316 expression_sql = f"{returning}{from_sql}{where_sql}{limit}" 1317 sql = f"UPDATE {this} SET {set_sql}{expression_sql}" 1318 return self.prepend_ctes(expression, sql)
1320 def values_sql(self, expression: exp.Values) -> str: 1321 # The VALUES clause is still valid in an `INSERT INTO ..` statement, for example 1322 if self.VALUES_AS_TABLE or not expression.find_ancestor(exp.From, exp.Join): 1323 args = self.expressions(expression) 1324 alias = self.sql(expression, "alias") 1325 values = f"VALUES{self.seg('')}{args}" 1326 values = ( 1327 f"({values})" 1328 if self.WRAP_DERIVED_VALUES and (alias or isinstance(expression.parent, exp.From)) 1329 else values 1330 ) 1331 return f"{values} AS {alias}" if alias else values 1332 1333 # Converts `VALUES...` expression into a series of select unions. 1334 # Note: If you have a lot of unions then this will result in a large number of recursive statements to 1335 # evaluate the expression. You may need to increase `sys.setrecursionlimit` to run and it can also be 1336 # very slow. 1337 expression = expression.copy() 1338 column_names = expression.alias and expression.args["alias"].columns 1339 1340 selects = [] 1341 1342 for i, tup in enumerate(expression.expressions): 1343 row = tup.expressions 1344 1345 if i == 0 and column_names: 1346 row = [ 1347 exp.alias_(value, column_name) for value, column_name in zip(row, column_names) 1348 ] 1349 1350 selects.append(exp.Select(expressions=row)) 1351 1352 subquery_expression: exp.Select | exp.Union = selects[0] 1353 if len(selects) > 1: 1354 for select in selects[1:]: 1355 subquery_expression = exp.union( 1356 subquery_expression, select, distinct=False, copy=False 1357 ) 1358 1359 return self.subquery_sql(subquery_expression.subquery(expression.alias, copy=False))
1372 def group_sql(self, expression: exp.Group) -> str: 1373 group_by = self.op_expressions("GROUP BY", expression) 1374 1375 if expression.args.get("all"): 1376 return f"{group_by} ALL" 1377 1378 grouping_sets = self.expressions(expression, key="grouping_sets", indent=False) 1379 grouping_sets = ( 1380 f"{self.seg('GROUPING SETS')} {self.wrap(grouping_sets)}" if grouping_sets else "" 1381 ) 1382 1383 cube = expression.args.get("cube", []) 1384 if seq_get(cube, 0) is True: 1385 return f"{group_by}{self.seg('WITH CUBE')}" 1386 else: 1387 cube_sql = self.expressions(expression, key="cube", indent=False) 1388 cube_sql = f"{self.seg('CUBE')} {self.wrap(cube_sql)}" if cube_sql else "" 1389 1390 rollup = expression.args.get("rollup", []) 1391 if seq_get(rollup, 0) is True: 1392 return f"{group_by}{self.seg('WITH ROLLUP')}" 1393 else: 1394 rollup_sql = self.expressions(expression, key="rollup", indent=False) 1395 rollup_sql = f"{self.seg('ROLLUP')} {self.wrap(rollup_sql)}" if rollup_sql else "" 1396 1397 groupings = csv( 1398 grouping_sets, 1399 cube_sql, 1400 rollup_sql, 1401 self.seg("WITH TOTALS") if expression.args.get("totals") else "", 1402 sep=self.GROUPINGS_SEP, 1403 ) 1404 1405 if expression.args.get("expressions") and groupings: 1406 group_by = f"{group_by}{self.GROUPINGS_SEP}" 1407 1408 return f"{group_by}{groupings}"
1414 def join_sql(self, expression: exp.Join) -> str: 1415 op_sql = " ".join( 1416 op 1417 for op in ( 1418 expression.method, 1419 "GLOBAL" if expression.args.get("global") else None, 1420 expression.side, 1421 expression.kind, 1422 expression.hint if self.JOIN_HINTS else None, 1423 ) 1424 if op 1425 ) 1426 on_sql = self.sql(expression, "on") 1427 using = expression.args.get("using") 1428 1429 if not on_sql and using: 1430 on_sql = csv(*(self.sql(column) for column in using)) 1431 1432 this_sql = self.sql(expression, "this") 1433 1434 if on_sql: 1435 on_sql = self.indent(on_sql, skip_first=True) 1436 space = self.seg(" " * self.pad) if self.pretty else " " 1437 if using: 1438 on_sql = f"{space}USING ({on_sql})" 1439 else: 1440 on_sql = f"{space}ON {on_sql}" 1441 elif not op_sql: 1442 return f", {this_sql}" 1443 1444 op_sql = f"{op_sql} JOIN" if op_sql else "JOIN" 1445 return f"{self.seg(op_sql)} {this_sql}{on_sql}"
1452 def lateral_sql(self, expression: exp.Lateral) -> str: 1453 this = self.sql(expression, "this") 1454 1455 if isinstance(expression.this, exp.Subquery): 1456 return f"LATERAL {this}" 1457 1458 if expression.args.get("view"): 1459 alias = expression.args["alias"] 1460 columns = self.expressions(alias, key="columns", flat=True) 1461 table = f" {alias.name}" if alias.name else "" 1462 columns = f" AS {columns}" if columns else "" 1463 op_sql = self.seg(f"LATERAL VIEW{' OUTER' if expression.args.get('outer') else ''}") 1464 return f"{op_sql}{self.sep()}{this}{table}{columns}" 1465 1466 alias = self.sql(expression, "alias") 1467 alias = f" AS {alias}" if alias else "" 1468 return f"LATERAL {this}{alias}"
1470 def limit_sql(self, expression: exp.Limit, top: bool = False) -> str: 1471 this = self.sql(expression, "this") 1472 args = ", ".join( 1473 sql 1474 for sql in ( 1475 self.sql(expression, "offset"), 1476 self.sql(expression, "expression"), 1477 ) 1478 if sql 1479 ) 1480 return f"{this}{self.seg('TOP' if top else 'LIMIT')} {args}"
1486 def setitem_sql(self, expression: exp.SetItem) -> str: 1487 kind = self.sql(expression, "kind") 1488 kind = f"{kind} " if kind else "" 1489 this = self.sql(expression, "this") 1490 expressions = self.expressions(expression) 1491 collate = self.sql(expression, "collate") 1492 collate = f" COLLATE {collate}" if collate else "" 1493 global_ = "GLOBAL " if expression.args.get("global") else "" 1494 return f"{global_}{kind}{this}{expressions}{collate}"
1496 def set_sql(self, expression: exp.Set) -> str: 1497 expressions = ( 1498 f" {self.expressions(expression, flat=True)}" if expression.expressions else "" 1499 ) 1500 tag = " TAG" if expression.args.get("tag") else "" 1501 return f"{'UNSET' if expression.args.get('unset') else 'SET'}{tag}{expressions}"
1506 def lock_sql(self, expression: exp.Lock) -> str: 1507 if not self.LOCKING_READS_SUPPORTED: 1508 self.unsupported("Locking reads using 'FOR UPDATE/SHARE' are not supported") 1509 return "" 1510 1511 lock_type = "FOR UPDATE" if expression.args["update"] else "FOR SHARE" 1512 expressions = self.expressions(expression, flat=True) 1513 expressions = f" OF {expressions}" if expressions else "" 1514 wait = expression.args.get("wait") 1515 1516 if wait is not None: 1517 if isinstance(wait, exp.Literal): 1518 wait = f" WAIT {self.sql(wait)}" 1519 else: 1520 wait = " NOWAIT" if wait else " SKIP LOCKED" 1521 1522 return f"{lock_type}{expressions}{wait or ''}"
1538 def loaddata_sql(self, expression: exp.LoadData) -> str: 1539 local = " LOCAL" if expression.args.get("local") else "" 1540 inpath = f" INPATH {self.sql(expression, 'inpath')}" 1541 overwrite = " OVERWRITE" if expression.args.get("overwrite") else "" 1542 this = f" INTO TABLE {self.sql(expression, 'this')}" 1543 partition = self.sql(expression, "partition") 1544 partition = f" {partition}" if partition else "" 1545 input_format = self.sql(expression, "input_format") 1546 input_format = f" INPUTFORMAT {input_format}" if input_format else "" 1547 serde = self.sql(expression, "serde") 1548 serde = f" SERDE {serde}" if serde else "" 1549 return f"LOAD DATA{local}{inpath}{overwrite}{this}{partition}{input_format}{serde}"
1571 def ordered_sql(self, expression: exp.Ordered) -> str: 1572 desc = expression.args.get("desc") 1573 asc = not desc 1574 1575 nulls_first = expression.args.get("nulls_first") 1576 nulls_last = not nulls_first 1577 nulls_are_large = self.NULL_ORDERING == "nulls_are_large" 1578 nulls_are_small = self.NULL_ORDERING == "nulls_are_small" 1579 nulls_are_last = self.NULL_ORDERING == "nulls_are_last" 1580 1581 sort_order = " DESC" if desc else "" 1582 nulls_sort_change = "" 1583 if nulls_first and ( 1584 (asc and nulls_are_large) or (desc and nulls_are_small) or nulls_are_last 1585 ): 1586 nulls_sort_change = " NULLS FIRST" 1587 elif ( 1588 nulls_last 1589 and ((asc and nulls_are_small) or (desc and nulls_are_large)) 1590 and not nulls_are_last 1591 ): 1592 nulls_sort_change = " NULLS LAST" 1593 1594 if nulls_sort_change and not self.NULL_ORDERING_SUPPORTED: 1595 self.unsupported( 1596 "Sorting in an ORDER BY on NULLS FIRST/NULLS LAST is not supported by this dialect" 1597 ) 1598 nulls_sort_change = "" 1599 1600 return f"{self.sql(expression, 'this')}{sort_order}{nulls_sort_change}"
1602 def matchrecognize_sql(self, expression: exp.MatchRecognize) -> str: 1603 partition = self.partition_by_sql(expression) 1604 order = self.sql(expression, "order") 1605 measures = self.expressions(expression, key="measures") 1606 measures = self.seg(f"MEASURES{self.seg(measures)}") if measures else "" 1607 rows = self.sql(expression, "rows") 1608 rows = self.seg(rows) if rows else "" 1609 after = self.sql(expression, "after") 1610 after = self.seg(after) if after else "" 1611 pattern = self.sql(expression, "pattern") 1612 pattern = self.seg(f"PATTERN ({pattern})") if pattern else "" 1613 definition_sqls = [ 1614 f"{self.sql(definition, 'alias')} AS {self.sql(definition, 'this')}" 1615 for definition in expression.args.get("define", []) 1616 ] 1617 definitions = self.expressions(sqls=definition_sqls) 1618 define = self.seg(f"DEFINE{self.seg(definitions)}") if definitions else "" 1619 body = "".join( 1620 ( 1621 partition, 1622 order, 1623 measures, 1624 rows, 1625 after, 1626 pattern, 1627 define, 1628 ) 1629 ) 1630 alias = self.sql(expression, "alias") 1631 alias = f" {alias}" if alias else "" 1632 return f"{self.seg('MATCH_RECOGNIZE')} {self.wrap(body)}{alias}"
1634 def query_modifiers(self, expression: exp.Expression, *sqls: str) -> str: 1635 limit: t.Optional[exp.Fetch | exp.Limit] = expression.args.get("limit") 1636 1637 # If the limit is generated as TOP, we need to ensure it's not generated twice 1638 with_offset_limit_modifiers = not isinstance(limit, exp.Limit) or not self.LIMIT_IS_TOP 1639 1640 if self.LIMIT_FETCH == "LIMIT" and isinstance(limit, exp.Fetch): 1641 limit = exp.Limit(expression=limit.args.get("count")) 1642 elif self.LIMIT_FETCH == "FETCH" and isinstance(limit, exp.Limit): 1643 limit = exp.Fetch(direction="FIRST", count=limit.expression) 1644 1645 fetch = isinstance(limit, exp.Fetch) 1646 1647 offset_limit_modifiers = ( 1648 self.offset_limit_modifiers(expression, fetch, limit) 1649 if with_offset_limit_modifiers 1650 else [] 1651 ) 1652 1653 return csv( 1654 *sqls, 1655 *[self.sql(join) for join in expression.args.get("joins") or []], 1656 self.sql(expression, "match"), 1657 *[self.sql(lateral) for lateral in expression.args.get("laterals") or []], 1658 self.sql(expression, "where"), 1659 self.sql(expression, "group"), 1660 self.sql(expression, "having"), 1661 *self.after_having_modifiers(expression), 1662 self.sql(expression, "order"), 1663 *offset_limit_modifiers, 1664 *self.after_limit_modifiers(expression), 1665 sep="", 1666 )
def
offset_limit_modifiers( self, expression: sqlglot.expressions.Expression, fetch: bool, limit: Union[sqlglot.expressions.Fetch, sqlglot.expressions.Limit, NoneType]) -> List[str]:
1668 def offset_limit_modifiers( 1669 self, expression: exp.Expression, fetch: bool, limit: t.Optional[exp.Fetch | exp.Limit] 1670 ) -> t.List[str]: 1671 return [ 1672 self.sql(expression, "offset") if fetch else self.sql(limit), 1673 self.sql(limit) if fetch else self.sql(expression, "offset"), 1674 ]
1676 def after_having_modifiers(self, expression: exp.Expression) -> t.List[str]: 1677 return [ 1678 self.sql(expression, "qualify"), 1679 self.seg("WINDOW ") + self.expressions(expression, key="windows", flat=True) 1680 if expression.args.get("windows") 1681 else "", 1682 self.sql(expression, "distribute"), 1683 self.sql(expression, "sort"), 1684 self.sql(expression, "cluster"), 1685 ]
1692 def select_sql(self, expression: exp.Select) -> str: 1693 hint = self.sql(expression, "hint") 1694 distinct = self.sql(expression, "distinct") 1695 distinct = f" {distinct}" if distinct else "" 1696 kind = self.sql(expression, "kind").upper() 1697 limit = expression.args.get("limit") 1698 top = ( 1699 self.limit_sql(limit, top=True) 1700 if isinstance(limit, exp.Limit) and self.LIMIT_IS_TOP 1701 else "" 1702 ) 1703 1704 expressions = self.expressions(expression) 1705 1706 if kind: 1707 if kind in self.SELECT_KINDS: 1708 kind = f" AS {kind}" 1709 else: 1710 if kind == "STRUCT": 1711 expressions = self.expressions( 1712 sqls=[ 1713 self.sql( 1714 exp.Struct( 1715 expressions=[ 1716 exp.column(e.output_name).eq( 1717 e.this if isinstance(e, exp.Alias) else e 1718 ) 1719 for e in expression.expressions 1720 ] 1721 ) 1722 ) 1723 ] 1724 ) 1725 kind = "" 1726 1727 expressions = f"{self.sep()}{expressions}" if expressions else expressions 1728 sql = self.query_modifiers( 1729 expression, 1730 f"SELECT{top}{hint}{distinct}{kind}{expressions}", 1731 self.sql(expression, "into", comment=False), 1732 self.sql(expression, "from", comment=False), 1733 ) 1734 return self.prepend_ctes(expression, sql)
1745 def star_sql(self, expression: exp.Star) -> str: 1746 except_ = self.expressions(expression, key="except", flat=True) 1747 except_ = f"{self.seg(self.STAR_MAPPING['except'])} ({except_})" if except_ else "" 1748 replace = self.expressions(expression, key="replace", flat=True) 1749 replace = f"{self.seg(self.STAR_MAPPING['replace'])} ({replace})" if replace else "" 1750 return f"*{except_}{replace}"
1767 def subquery_sql(self, expression: exp.Subquery, sep: str = " AS ") -> str: 1768 alias = self.sql(expression, "alias") 1769 alias = f"{sep}{alias}" if alias else "" 1770 1771 pivots = self.expressions(expression, key="pivots", sep=" ", flat=True) 1772 pivots = f" {pivots}" if pivots else "" 1773 1774 sql = self.query_modifiers(expression, self.wrap(expression), alias, pivots) 1775 return self.prepend_ctes(expression, sql)
1792 def unnest_sql(self, expression: exp.Unnest) -> str: 1793 args = self.expressions(expression, flat=True) 1794 alias = expression.args.get("alias") 1795 if alias and self.UNNEST_COLUMN_ONLY: 1796 columns = alias.columns 1797 alias = self.sql(columns[0]) if columns else "" 1798 else: 1799 alias = self.sql(expression, "alias") 1800 alias = f" AS {alias}" if alias else alias 1801 ordinality = " WITH ORDINALITY" if expression.args.get("ordinality") else "" 1802 offset = expression.args.get("offset") 1803 offset = f" WITH OFFSET AS {self.sql(offset)}" if offset else "" 1804 return f"UNNEST({args}){ordinality}{alias}{offset}"
1810 def window_sql(self, expression: exp.Window) -> str: 1811 this = self.sql(expression, "this") 1812 partition = self.partition_by_sql(expression) 1813 order = expression.args.get("order") 1814 order = self.order_sql(order, flat=True) if order else "" 1815 spec = self.sql(expression, "spec") 1816 alias = self.sql(expression, "alias") 1817 over = self.sql(expression, "over") or "OVER" 1818 1819 this = f"{this} {'AS' if expression.arg_key == 'windows' else over}" 1820 1821 first = expression.args.get("first") 1822 if first is None: 1823 first = "" 1824 else: 1825 first = "FIRST" if first else "LAST" 1826 1827 if not partition and not order and not spec and alias: 1828 return f"{this} {alias}" 1829 1830 args = " ".join(arg for arg in (alias, first, partition, order, spec) if arg) 1831 return f"{this} ({args})"
def
partition_by_sql( self, expression: sqlglot.expressions.Window | sqlglot.expressions.MatchRecognize) -> str:
1837 def windowspec_sql(self, expression: exp.WindowSpec) -> str: 1838 kind = self.sql(expression, "kind") 1839 start = csv(self.sql(expression, "start"), self.sql(expression, "start_side"), sep=" ") 1840 end = ( 1841 csv(self.sql(expression, "end"), self.sql(expression, "end_side"), sep=" ") 1842 or "CURRENT ROW" 1843 ) 1844 return f"{kind} BETWEEN {start} AND {end}"
1878 def case_sql(self, expression: exp.Case) -> str: 1879 this = self.sql(expression, "this") 1880 statements = [f"CASE {this}" if this else "CASE"] 1881 1882 for e in expression.args["ifs"]: 1883 statements.append(f"WHEN {self.sql(e, 'this')}") 1884 statements.append(f"THEN {self.sql(e, 'true')}") 1885 1886 default = self.sql(expression, "default") 1887 1888 if default: 1889 statements.append(f"ELSE {default}") 1890 1891 statements.append("END") 1892 1893 if self.pretty and self.text_width(statements) > self.max_text_width: 1894 return self.indent("\n".join(statements), skip_first=True, skip_last=True) 1895 1896 return " ".join(statements)
1913 def trim_sql(self, expression: exp.Trim) -> str: 1914 trim_type = self.sql(expression, "position") 1915 1916 if trim_type == "LEADING": 1917 return self.func("LTRIM", expression.this) 1918 elif trim_type == "TRAILING": 1919 return self.func("RTRIM", expression.this) 1920 else: 1921 return self.func("TRIM", expression.this, expression.expression)
1933 def foreignkey_sql(self, expression: exp.ForeignKey) -> str: 1934 expressions = self.expressions(expression, flat=True) 1935 reference = self.sql(expression, "reference") 1936 reference = f" {reference}" if reference else "" 1937 delete = self.sql(expression, "delete") 1938 delete = f" ON DELETE {delete}" if delete else "" 1939 update = self.sql(expression, "update") 1940 update = f" ON UPDATE {update}" if update else "" 1941 return f"FOREIGN KEY ({expressions}){reference}{delete}{update}"
1943 def primarykey_sql(self, expression: exp.ForeignKey) -> str: 1944 expressions = self.expressions(expression, flat=True) 1945 options = self.expressions(expression, key="options", flat=True, sep=" ") 1946 options = f" {options}" if options else "" 1947 return f"PRIMARY KEY ({expressions}){options}"
1960 def jsonobject_sql(self, expression: exp.JSONObject) -> str: 1961 null_handling = expression.args.get("null_handling") 1962 null_handling = f" {null_handling}" if null_handling else "" 1963 unique_keys = expression.args.get("unique_keys") 1964 if unique_keys is not None: 1965 unique_keys = f" {'WITH' if unique_keys else 'WITHOUT'} UNIQUE KEYS" 1966 else: 1967 unique_keys = "" 1968 return_type = self.sql(expression, "return_type") 1969 return_type = f" RETURNING {return_type}" if return_type else "" 1970 format_json = " FORMAT JSON" if expression.args.get("format_json") else "" 1971 encoding = self.sql(expression, "encoding") 1972 encoding = f" ENCODING {encoding}" if encoding else "" 1973 return self.func( 1974 "JSON_OBJECT", 1975 *expression.expressions, 1976 suffix=f"{null_handling}{unique_keys}{return_type}{format_json}{encoding})", 1977 )
1979 def openjsoncolumndef_sql(self, expression: exp.OpenJSONColumnDef) -> str: 1980 this = self.sql(expression, "this") 1981 kind = self.sql(expression, "kind") 1982 path = self.sql(expression, "path") 1983 path = f" {path}" if path else "" 1984 as_json = " AS JSON" if expression.args.get("as_json") else "" 1985 return f"{this} {kind}{path}{as_json}"
1987 def openjson_sql(self, expression: exp.OpenJSON) -> str: 1988 this = self.sql(expression, "this") 1989 path = self.sql(expression, "path") 1990 path = f", {path}" if path else "" 1991 expressions = self.expressions(expression) 1992 with_ = ( 1993 f" WITH ({self.seg(self.indent(expressions), sep='')}{self.seg(')', sep='')}" 1994 if expressions 1995 else "" 1996 ) 1997 return f"OPENJSON({this}{path}){with_}"
1999 def in_sql(self, expression: exp.In) -> str: 2000 query = expression.args.get("query") 2001 unnest = expression.args.get("unnest") 2002 field = expression.args.get("field") 2003 is_global = " GLOBAL" if expression.args.get("is_global") else "" 2004 2005 if query: 2006 in_sql = self.wrap(query) 2007 elif unnest: 2008 in_sql = self.in_unnest_op(unnest) 2009 elif field: 2010 in_sql = self.sql(field) 2011 else: 2012 in_sql = f"({self.expressions(expression, flat=True)})" 2013 2014 return f"{self.sql(expression, 'this')}{is_global} IN {in_sql}"
2019 def interval_sql(self, expression: exp.Interval) -> str: 2020 unit = self.sql(expression, "unit") 2021 if not self.INTERVAL_ALLOWS_PLURAL_FORM: 2022 unit = self.TIME_PART_SINGULARS.get(unit.lower(), unit) 2023 unit = f" {unit}" if unit else "" 2024 2025 if self.SINGLE_STRING_INTERVAL: 2026 this = expression.this.name if expression.this else "" 2027 return f"INTERVAL '{this}{unit}'" if this else f"INTERVAL{unit}" 2028 2029 this = self.sql(expression, "this") 2030 if this: 2031 unwrapped = isinstance(expression.this, self.UNWRAPPED_INTERVAL_VALUES) 2032 this = f" {this}" if unwrapped else f" ({this})" 2033 2034 return f"INTERVAL{this}{unit}"
2039 def reference_sql(self, expression: exp.Reference) -> str: 2040 this = self.sql(expression, "this") 2041 expressions = self.expressions(expression, flat=True) 2042 expressions = f"({expressions})" if expressions else "" 2043 options = self.expressions(expression, key="options", flat=True, sep=" ") 2044 options = f" {options}" if options else "" 2045 return f"REFERENCES {this}{expressions}{options}"
2050 def paren_sql(self, expression: exp.Paren) -> str: 2051 if isinstance(expression.unnest(), exp.Select): 2052 sql = self.wrap(expression) 2053 else: 2054 sql = self.seg(self.indent(self.sql(expression, "this")), sep="") 2055 sql = f"({sql}{self.seg(')', sep='')}" 2056 2057 return self.prepend_ctes(expression, sql)
2090 def connector_sql(self, expression: exp.Connector, op: str) -> str: 2091 if not self.pretty: 2092 return self.binary(expression, op) 2093 2094 sqls = tuple( 2095 self.maybe_comment(self.sql(e), e, e.parent.comments or []) if i != 1 else self.sql(e) 2096 for i, e in enumerate(expression.flatten(unnest=False)) 2097 ) 2098 2099 sep = "\n" if self.text_width(sqls) > self.max_text_width else " " 2100 return f"{sep}{op} ".join(sqls)
def
cast_sql( self, expression: sqlglot.expressions.Cast, safe_prefix: Optional[str] = None) -> str:
2120 def cast_sql(self, expression: exp.Cast, safe_prefix: t.Optional[str] = None) -> str: 2121 format_sql = self.sql(expression, "format") 2122 format_sql = f" FORMAT {format_sql}" if format_sql else "" 2123 return f"{safe_prefix or ''}CAST({self.sql(expression, 'this')} AS {self.sql(expression, 'to')}{format_sql})"
2135 def comment_sql(self, expression: exp.Comment) -> str: 2136 this = self.sql(expression, "this") 2137 kind = expression.args["kind"] 2138 exists_sql = " IF EXISTS " if expression.args.get("exists") else " " 2139 expression_sql = self.sql(expression, "expression") 2140 return f"COMMENT{exists_sql}ON {kind} {this} IS {expression_sql}"
2142 def mergetreettlaction_sql(self, expression: exp.MergeTreeTTLAction) -> str: 2143 this = self.sql(expression, "this") 2144 delete = " DELETE" if expression.args.get("delete") else "" 2145 recompress = self.sql(expression, "recompress") 2146 recompress = f" RECOMPRESS {recompress}" if recompress else "" 2147 to_disk = self.sql(expression, "to_disk") 2148 to_disk = f" TO DISK {to_disk}" if to_disk else "" 2149 to_volume = self.sql(expression, "to_volume") 2150 to_volume = f" TO VOLUME {to_volume}" if to_volume else "" 2151 return f"{this}{delete}{recompress}{to_disk}{to_volume}"
2153 def mergetreettl_sql(self, expression: exp.MergeTreeTTL) -> str: 2154 where = self.sql(expression, "where") 2155 group = self.sql(expression, "group") 2156 aggregates = self.expressions(expression, key="aggregates") 2157 aggregates = self.seg("SET") + self.seg(aggregates) if aggregates else "" 2158 2159 if not (where or group or aggregates) and len(expression.expressions) == 1: 2160 return f"TTL {self.expressions(expression, flat=True)}" 2161 2162 return f"TTL{self.seg(self.expressions(expression))}{where}{group}{aggregates}"
2179 def altercolumn_sql(self, expression: exp.AlterColumn) -> str: 2180 this = self.sql(expression, "this") 2181 2182 dtype = self.sql(expression, "dtype") 2183 if dtype: 2184 collate = self.sql(expression, "collate") 2185 collate = f" COLLATE {collate}" if collate else "" 2186 using = self.sql(expression, "using") 2187 using = f" USING {using}" if using else "" 2188 return f"ALTER COLUMN {this} TYPE {dtype}{collate}{using}" 2189 2190 default = self.sql(expression, "default") 2191 if default: 2192 return f"ALTER COLUMN {this} SET DEFAULT {default}" 2193 2194 if not expression.args.get("drop"): 2195 self.unsupported("Unsupported ALTER COLUMN syntax") 2196 2197 return f"ALTER COLUMN {this} DROP DEFAULT"
2199 def renametable_sql(self, expression: exp.RenameTable) -> str: 2200 if not self.RENAME_TABLE_WITH_DB: 2201 # Remove db from tables 2202 expression = expression.transform( 2203 lambda n: exp.table_(n.this) if isinstance(n, exp.Table) else n 2204 ) 2205 this = self.sql(expression, "this") 2206 return f"RENAME TO {this}"
2208 def altertable_sql(self, expression: exp.AlterTable) -> str: 2209 actions = expression.args["actions"] 2210 2211 if isinstance(actions[0], exp.ColumnDef): 2212 actions = self.expressions(expression, key="actions", prefix="ADD COLUMN ") 2213 elif isinstance(actions[0], exp.Schema): 2214 actions = self.expressions(expression, key="actions", prefix="ADD COLUMNS ") 2215 elif isinstance(actions[0], exp.Delete): 2216 actions = self.expressions(expression, key="actions", flat=True) 2217 else: 2218 actions = self.expressions(expression, key="actions") 2219 2220 exists = " IF EXISTS" if expression.args.get("exists") else "" 2221 return f"ALTER TABLE{exists} {self.sql(expression, 'this')} {actions}"
2228 def addconstraint_sql(self, expression: exp.AddConstraint) -> str: 2229 this = self.sql(expression, "this") 2230 expression_ = self.sql(expression, "expression") 2231 add_constraint = f"ADD CONSTRAINT {this}" if this else "ADD" 2232 2233 enforced = expression.args.get("enforced") 2234 if enforced is not None: 2235 return f"{add_constraint} CHECK ({expression_}){' ENFORCED' if enforced else ''}" 2236 2237 return f"{add_constraint} {expression_}"
2362 def function_fallback_sql(self, expression: exp.Func) -> str: 2363 args = [] 2364 2365 for key in expression.arg_types: 2366 arg_value = expression.args.get(key) 2367 2368 if isinstance(arg_value, list): 2369 for value in arg_value: 2370 args.append(value) 2371 elif arg_value is not None: 2372 args.append(arg_value) 2373 2374 return self.func(expression.sql_name(), *args)
def
func( self, name: str, *args: Union[str, sqlglot.expressions.Expression, NoneType], prefix: str = '(', suffix: str = ')') -> str:
2385 def format_args(self, *args: t.Optional[str | exp.Expression]) -> str: 2386 arg_sqls = tuple(self.sql(arg) for arg in args if arg is not None) 2387 if self.pretty and self.text_width(arg_sqls) > self.max_text_width: 2388 return self.indent("\n" + f",\n".join(arg_sqls) + "\n", skip_first=True, skip_last=True) 2389 return ", ".join(arg_sqls)
def
expressions( self, expression: Optional[sqlglot.expressions.Expression] = None, key: Optional[str] = None, sqls: Optional[List[str]] = None, flat: bool = False, indent: bool = True, skip_first: bool = False, sep: str = ', ', prefix: str = '') -> str:
2399 def expressions( 2400 self, 2401 expression: t.Optional[exp.Expression] = None, 2402 key: t.Optional[str] = None, 2403 sqls: t.Optional[t.List[str]] = None, 2404 flat: bool = False, 2405 indent: bool = True, 2406 skip_first: bool = False, 2407 sep: str = ", ", 2408 prefix: str = "", 2409 ) -> str: 2410 expressions = expression.args.get(key or "expressions") if expression else sqls 2411 2412 if not expressions: 2413 return "" 2414 2415 if flat: 2416 return sep.join(self.sql(e) for e in expressions) 2417 2418 num_sqls = len(expressions) 2419 2420 # These are calculated once in case we have the leading_comma / pretty option set, correspondingly 2421 pad = " " * self.pad 2422 stripped_sep = sep.strip() 2423 2424 result_sqls = [] 2425 for i, e in enumerate(expressions): 2426 sql = self.sql(e, comment=False) 2427 comments = self.maybe_comment("", e) if isinstance(e, exp.Expression) else "" 2428 2429 if self.pretty: 2430 if self.leading_comma: 2431 result_sqls.append(f"{sep if i > 0 else pad}{prefix}{sql}{comments}") 2432 else: 2433 result_sqls.append( 2434 f"{prefix}{sql}{stripped_sep if i + 1 < num_sqls else ''}{comments}" 2435 ) 2436 else: 2437 result_sqls.append(f"{prefix}{sql}{comments}{sep if i + 1 < num_sqls else ''}") 2438 2439 result_sql = "\n".join(result_sqls) if self.pretty else "".join(result_sqls) 2440 return self.indent(result_sql, skip_first=skip_first) if indent else result_sql
def
op_expressions( self, op: str, expression: sqlglot.expressions.Expression, flat: bool = False) -> str:
2442 def op_expressions(self, op: str, expression: exp.Expression, flat: bool = False) -> str: 2443 flat = flat or isinstance(expression.parent, exp.Properties) 2444 expressions_sql = self.expressions(expression, flat=flat) 2445 if flat: 2446 return f"{op} {expressions_sql}" 2447 return f"{self.seg(op)}{self.sep() if expressions_sql else ''}{expressions_sql}"
2449 def naked_property(self, expression: exp.Property) -> str: 2450 property_name = exp.Properties.PROPERTY_TO_NAME.get(expression.__class__) 2451 if not property_name: 2452 self.unsupported(f"Unsupported property {expression.__class__.__name__}") 2453 return f"{property_name} {self.sql(expression, 'this')}"
2468 def userdefinedfunction_sql(self, expression: exp.UserDefinedFunction) -> str: 2469 this = self.sql(expression, "this") 2470 expressions = self.no_identify(self.expressions, expression) 2471 expressions = ( 2472 self.wrap(expressions) if expression.args.get("wrapped") else f" {expressions}" 2473 ) 2474 return f"{this}{expressions}"
2484 def when_sql(self, expression: exp.When) -> str: 2485 matched = "MATCHED" if expression.args["matched"] else "NOT MATCHED" 2486 source = " BY SOURCE" if self.MATCHED_BY_SOURCE and expression.args.get("source") else "" 2487 condition = self.sql(expression, "condition") 2488 condition = f" AND {condition}" if condition else "" 2489 2490 then_expression = expression.args.get("then") 2491 if isinstance(then_expression, exp.Insert): 2492 then = f"INSERT {self.sql(then_expression, 'this')}" 2493 if "expression" in then_expression.args: 2494 then += f" VALUES {self.sql(then_expression, 'expression')}" 2495 elif isinstance(then_expression, exp.Update): 2496 if isinstance(then_expression.args.get("expressions"), exp.Star): 2497 then = f"UPDATE {self.sql(then_expression, 'expressions')}" 2498 else: 2499 then = f"UPDATE SET {self.expressions(then_expression, flat=True)}" 2500 else: 2501 then = self.sql(then_expression) 2502 return f"WHEN {matched}{source}{condition} THEN {then}"
2516 def dictproperty_sql(self, expression: exp.DictProperty) -> str: 2517 this = self.sql(expression, "this") 2518 kind = self.sql(expression, "kind") 2519 settings_sql = self.expressions(expression, key="settings", sep=" ") 2520 args = f"({self.sep('')}{settings_sql}{self.seg(')', sep='')}" if settings_sql else "()" 2521 return f"{this}({kind}{args})"
2535 def clusteredbyproperty_sql(self, expression: exp.ClusteredByProperty) -> str: 2536 expressions = self.expressions(expression, key="expressions", flat=True) 2537 sorted_by = self.expressions(expression, key="sorted_by", flat=True) 2538 sorted_by = f" SORTED BY ({sorted_by})" if sorted_by else "" 2539 buckets = self.sql(expression, "buckets") 2540 return f"CLUSTERED BY ({expressions}){sorted_by} INTO {buckets} BUCKETS"
2542 def anyvalue_sql(self, expression: exp.AnyValue) -> str: 2543 this = self.sql(expression, "this") 2544 having = self.sql(expression, "having") 2545 2546 if having: 2547 this = f"{this} HAVING {'MAX' if expression.args.get('max') else 'MIN'} {having}" 2548 2549 return self.func("ANY_VALUE", this)
2551 def querytransform_sql(self, expression: exp.QueryTransform) -> str: 2552 transform = self.func("TRANSFORM", *expression.expressions) 2553 row_format_before = self.sql(expression, "row_format_before") 2554 row_format_before = f" {row_format_before}" if row_format_before else "" 2555 record_writer = self.sql(expression, "record_writer") 2556 record_writer = f" RECORDWRITER {record_writer}" if record_writer else "" 2557 using = f" USING {self.sql(expression, 'command_script')}" 2558 schema = self.sql(expression, "schema") 2559 schema = f" AS {schema}" if schema else "" 2560 row_format_after = self.sql(expression, "row_format_after") 2561 row_format_after = f" {row_format_after}" if row_format_after else "" 2562 record_reader = self.sql(expression, "record_reader") 2563 record_reader = f" RECORDREADER {record_reader}" if record_reader else "" 2564 return f"{transform}{row_format_before}{record_writer}{using}{schema}{row_format_after}{record_reader}"
def
cached_generator( cache: Optional[Dict[int, str]] = None) -> Callable[[sqlglot.expressions.Expression], str]:
2567def cached_generator( 2568 cache: t.Optional[t.Dict[int, str]] = None 2569) -> t.Callable[[exp.Expression], str]: 2570 """Returns a cached generator.""" 2571 cache = {} if cache is None else cache 2572 generator = Generator(normalize=True, identify="safe") 2573 return lambda e: generator.generate(e, cache)
Returns a cached generator.