sqlglot.dialects.presto
1from __future__ import annotations 2 3import typing as t 4 5from sqlglot import exp, generator, parser, tokens, transforms 6from sqlglot.dialects.dialect import ( 7 Dialect, 8 NormalizationStrategy, 9 binary_from_function, 10 bool_xor_sql, 11 date_trunc_to_time, 12 datestrtodate_sql, 13 encode_decode_sql, 14 build_formatted_time, 15 if_sql, 16 left_to_substring_sql, 17 no_ilike_sql, 18 no_pivot_sql, 19 no_safe_divide_sql, 20 no_timestamp_sql, 21 regexp_extract_sql, 22 rename_func, 23 right_to_substring_sql, 24 struct_extract_sql, 25 timestamptrunc_sql, 26 timestrtotime_sql, 27 ts_or_ds_add_cast, 28) 29from sqlglot.dialects.mysql import MySQL 30from sqlglot.helper import apply_index_offset, seq_get 31from sqlglot.tokens import TokenType 32 33 34def _explode_to_unnest_sql(self: Presto.Generator, expression: exp.Lateral) -> str: 35 if isinstance(expression.this, exp.Explode): 36 return self.sql( 37 exp.Join( 38 this=exp.Unnest( 39 expressions=[expression.this.this], 40 alias=expression.args.get("alias"), 41 offset=isinstance(expression.this, exp.Posexplode), 42 ), 43 kind="cross", 44 ) 45 ) 46 return self.lateral_sql(expression) 47 48 49def _initcap_sql(self: Presto.Generator, expression: exp.Initcap) -> str: 50 regex = r"(\w)(\w*)" 51 return f"REGEXP_REPLACE({self.sql(expression, 'this')}, '{regex}', x -> UPPER(x[1]) || LOWER(x[2]))" 52 53 54def _no_sort_array(self: Presto.Generator, expression: exp.SortArray) -> str: 55 if expression.args.get("asc") == exp.false(): 56 comparator = "(a, b) -> CASE WHEN a < b THEN 1 WHEN a > b THEN -1 ELSE 0 END" 57 else: 58 comparator = None 59 return self.func("ARRAY_SORT", expression.this, comparator) 60 61 62def _schema_sql(self: Presto.Generator, expression: exp.Schema) -> str: 63 if isinstance(expression.parent, exp.Property): 64 columns = ", ".join(f"'{c.name}'" for c in expression.expressions) 65 return f"ARRAY[{columns}]" 66 67 if expression.parent: 68 for schema in expression.parent.find_all(exp.Schema): 69 column_defs = schema.find_all(exp.ColumnDef) 70 if column_defs and isinstance(schema.parent, exp.Property): 71 expression.expressions.extend(column_defs) 72 73 return self.schema_sql(expression) 74 75 76def _quantile_sql(self: Presto.Generator, expression: exp.Quantile) -> str: 77 self.unsupported("Presto does not support exact quantiles") 78 return self.func("APPROX_PERCENTILE", expression.this, expression.args.get("quantile")) 79 80 81def _str_to_time_sql( 82 self: Presto.Generator, expression: exp.StrToDate | exp.StrToTime | exp.TsOrDsToDate 83) -> str: 84 return self.func("DATE_PARSE", expression.this, self.format_time(expression)) 85 86 87def _ts_or_ds_to_date_sql(self: Presto.Generator, expression: exp.TsOrDsToDate) -> str: 88 time_format = self.format_time(expression) 89 if time_format and time_format not in (Presto.TIME_FORMAT, Presto.DATE_FORMAT): 90 return self.sql(exp.cast(_str_to_time_sql(self, expression), "DATE")) 91 return self.sql(exp.cast(exp.cast(expression.this, "TIMESTAMP"), "DATE")) 92 93 94def _ts_or_ds_add_sql(self: Presto.Generator, expression: exp.TsOrDsAdd) -> str: 95 expression = ts_or_ds_add_cast(expression) 96 unit = exp.Literal.string(expression.text("unit") or "DAY") 97 return self.func("DATE_ADD", unit, expression.expression, expression.this) 98 99 100def _ts_or_ds_diff_sql(self: Presto.Generator, expression: exp.TsOrDsDiff) -> str: 101 this = exp.cast(expression.this, "TIMESTAMP") 102 expr = exp.cast(expression.expression, "TIMESTAMP") 103 unit = exp.Literal.string(expression.text("unit") or "DAY") 104 return self.func("DATE_DIFF", unit, expr, this) 105 106 107def _build_approx_percentile(args: t.List) -> exp.Expression: 108 if len(args) == 4: 109 return exp.ApproxQuantile( 110 this=seq_get(args, 0), 111 weight=seq_get(args, 1), 112 quantile=seq_get(args, 2), 113 accuracy=seq_get(args, 3), 114 ) 115 if len(args) == 3: 116 return exp.ApproxQuantile( 117 this=seq_get(args, 0), quantile=seq_get(args, 1), accuracy=seq_get(args, 2) 118 ) 119 return exp.ApproxQuantile.from_arg_list(args) 120 121 122def _build_from_unixtime(args: t.List) -> exp.Expression: 123 if len(args) == 3: 124 return exp.UnixToTime( 125 this=seq_get(args, 0), 126 hours=seq_get(args, 1), 127 minutes=seq_get(args, 2), 128 ) 129 if len(args) == 2: 130 return exp.UnixToTime(this=seq_get(args, 0), zone=seq_get(args, 1)) 131 132 return exp.UnixToTime.from_arg_list(args) 133 134 135def _unnest_sequence(expression: exp.Expression) -> exp.Expression: 136 if isinstance(expression, exp.Table): 137 if isinstance(expression.this, exp.GenerateSeries): 138 unnest = exp.Unnest(expressions=[expression.this]) 139 140 if expression.alias: 141 return exp.alias_(unnest, alias="_u", table=[expression.alias], copy=False) 142 return unnest 143 return expression 144 145 146def _first_last_sql(self: Presto.Generator, expression: exp.Func) -> str: 147 """ 148 Trino doesn't support FIRST / LAST as functions, but they're valid in the context 149 of MATCH_RECOGNIZE, so we need to preserve them in that case. In all other cases 150 they're converted into an ARBITRARY call. 151 152 Reference: https://trino.io/docs/current/sql/match-recognize.html#logical-navigation-functions 153 """ 154 if isinstance(expression.find_ancestor(exp.MatchRecognize, exp.Select), exp.MatchRecognize): 155 return self.function_fallback_sql(expression) 156 157 return rename_func("ARBITRARY")(self, expression) 158 159 160def _unix_to_time_sql(self: Presto.Generator, expression: exp.UnixToTime) -> str: 161 scale = expression.args.get("scale") 162 timestamp = self.sql(expression, "this") 163 if scale in (None, exp.UnixToTime.SECONDS): 164 return rename_func("FROM_UNIXTIME")(self, expression) 165 166 return f"FROM_UNIXTIME(CAST({timestamp} AS DOUBLE) / POW(10, {scale}))" 167 168 169def _to_int(expression: exp.Expression) -> exp.Expression: 170 if not expression.type: 171 from sqlglot.optimizer.annotate_types import annotate_types 172 173 annotate_types(expression) 174 if expression.type and expression.type.this not in exp.DataType.INTEGER_TYPES: 175 return exp.cast(expression, to=exp.DataType.Type.BIGINT) 176 return expression 177 178 179def _build_to_char(args: t.List) -> exp.TimeToStr: 180 fmt = seq_get(args, 1) 181 if isinstance(fmt, exp.Literal): 182 # We uppercase this to match Teradata's format mapping keys 183 fmt.set("this", fmt.this.upper()) 184 185 # We use "teradata" on purpose here, because the time formats are different in Presto. 186 # See https://prestodb.io/docs/current/functions/teradata.html?highlight=to_char#to_char 187 return build_formatted_time(exp.TimeToStr, "teradata")(args) 188 189 190class Presto(Dialect): 191 INDEX_OFFSET = 1 192 NULL_ORDERING = "nulls_are_last" 193 TIME_FORMAT = MySQL.TIME_FORMAT 194 TIME_MAPPING = MySQL.TIME_MAPPING 195 STRICT_STRING_CONCAT = True 196 SUPPORTS_SEMI_ANTI_JOIN = False 197 TYPED_DIVISION = True 198 TABLESAMPLE_SIZE_IS_PERCENT = True 199 200 # https://github.com/trinodb/trino/issues/17 201 # https://github.com/trinodb/trino/issues/12289 202 # https://github.com/prestodb/presto/issues/2863 203 NORMALIZATION_STRATEGY = NormalizationStrategy.CASE_INSENSITIVE 204 205 class Tokenizer(tokens.Tokenizer): 206 UNICODE_STRINGS = [ 207 (prefix + q, q) 208 for q in t.cast(t.List[str], tokens.Tokenizer.QUOTES) 209 for prefix in ("U&", "u&") 210 ] 211 212 KEYWORDS = { 213 **tokens.Tokenizer.KEYWORDS, 214 "START": TokenType.BEGIN, 215 "MATCH_RECOGNIZE": TokenType.MATCH_RECOGNIZE, 216 "ROW": TokenType.STRUCT, 217 "IPADDRESS": TokenType.IPADDRESS, 218 "IPPREFIX": TokenType.IPPREFIX, 219 } 220 221 class Parser(parser.Parser): 222 VALUES_FOLLOWED_BY_PAREN = False 223 224 FUNCTIONS = { 225 **parser.Parser.FUNCTIONS, 226 "ARBITRARY": exp.AnyValue.from_arg_list, 227 "APPROX_DISTINCT": exp.ApproxDistinct.from_arg_list, 228 "APPROX_PERCENTILE": _build_approx_percentile, 229 "BITWISE_AND": binary_from_function(exp.BitwiseAnd), 230 "BITWISE_NOT": lambda args: exp.BitwiseNot(this=seq_get(args, 0)), 231 "BITWISE_OR": binary_from_function(exp.BitwiseOr), 232 "BITWISE_XOR": binary_from_function(exp.BitwiseXor), 233 "CARDINALITY": exp.ArraySize.from_arg_list, 234 "CONTAINS": exp.ArrayContains.from_arg_list, 235 "DATE_ADD": lambda args: exp.DateAdd( 236 this=seq_get(args, 2), expression=seq_get(args, 1), unit=seq_get(args, 0) 237 ), 238 "DATE_DIFF": lambda args: exp.DateDiff( 239 this=seq_get(args, 2), expression=seq_get(args, 1), unit=seq_get(args, 0) 240 ), 241 "DATE_FORMAT": build_formatted_time(exp.TimeToStr, "presto"), 242 "DATE_PARSE": build_formatted_time(exp.StrToTime, "presto"), 243 "DATE_TRUNC": date_trunc_to_time, 244 "ELEMENT_AT": lambda args: exp.Bracket( 245 this=seq_get(args, 0), expressions=[seq_get(args, 1)], offset=1, safe=True 246 ), 247 "FROM_HEX": exp.Unhex.from_arg_list, 248 "FROM_UNIXTIME": _build_from_unixtime, 249 "FROM_UTF8": lambda args: exp.Decode( 250 this=seq_get(args, 0), replace=seq_get(args, 1), charset=exp.Literal.string("utf-8") 251 ), 252 "NOW": exp.CurrentTimestamp.from_arg_list, 253 "REGEXP_EXTRACT": lambda args: exp.RegexpExtract( 254 this=seq_get(args, 0), expression=seq_get(args, 1), group=seq_get(args, 2) 255 ), 256 "REGEXP_REPLACE": lambda args: exp.RegexpReplace( 257 this=seq_get(args, 0), 258 expression=seq_get(args, 1), 259 replacement=seq_get(args, 2) or exp.Literal.string(""), 260 ), 261 "ROW": exp.Struct.from_arg_list, 262 "SEQUENCE": exp.GenerateSeries.from_arg_list, 263 "SET_AGG": exp.ArrayUniqueAgg.from_arg_list, 264 "SPLIT_TO_MAP": exp.StrToMap.from_arg_list, 265 "STRPOS": lambda args: exp.StrPosition( 266 this=seq_get(args, 0), substr=seq_get(args, 1), instance=seq_get(args, 2) 267 ), 268 "TO_CHAR": _build_to_char, 269 "TO_HEX": exp.Hex.from_arg_list, 270 "TO_UNIXTIME": exp.TimeToUnix.from_arg_list, 271 "TO_UTF8": lambda args: exp.Encode( 272 this=seq_get(args, 0), charset=exp.Literal.string("utf-8") 273 ), 274 } 275 276 FUNCTION_PARSERS = parser.Parser.FUNCTION_PARSERS.copy() 277 FUNCTION_PARSERS.pop("TRIM") 278 279 class Generator(generator.Generator): 280 INTERVAL_ALLOWS_PLURAL_FORM = False 281 JOIN_HINTS = False 282 TABLE_HINTS = False 283 QUERY_HINTS = False 284 IS_BOOL_ALLOWED = False 285 TZ_TO_WITH_TIME_ZONE = True 286 NVL2_SUPPORTED = False 287 STRUCT_DELIMITER = ("(", ")") 288 LIMIT_ONLY_LITERALS = True 289 SUPPORTS_SINGLE_ARG_CONCAT = False 290 LIKE_PROPERTY_INSIDE_SCHEMA = True 291 MULTI_ARG_DISTINCT = False 292 293 PROPERTIES_LOCATION = { 294 **generator.Generator.PROPERTIES_LOCATION, 295 exp.LocationProperty: exp.Properties.Location.UNSUPPORTED, 296 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 297 } 298 299 TYPE_MAPPING = { 300 **generator.Generator.TYPE_MAPPING, 301 exp.DataType.Type.INT: "INTEGER", 302 exp.DataType.Type.FLOAT: "REAL", 303 exp.DataType.Type.BINARY: "VARBINARY", 304 exp.DataType.Type.TEXT: "VARCHAR", 305 exp.DataType.Type.TIMETZ: "TIME", 306 exp.DataType.Type.TIMESTAMPTZ: "TIMESTAMP", 307 exp.DataType.Type.STRUCT: "ROW", 308 exp.DataType.Type.DATETIME: "TIMESTAMP", 309 exp.DataType.Type.DATETIME64: "TIMESTAMP", 310 } 311 312 TRANSFORMS = { 313 **generator.Generator.TRANSFORMS, 314 exp.AnyValue: rename_func("ARBITRARY"), 315 exp.ApproxDistinct: lambda self, e: self.func( 316 "APPROX_DISTINCT", e.this, e.args.get("accuracy") 317 ), 318 exp.ApproxQuantile: rename_func("APPROX_PERCENTILE"), 319 exp.ArgMax: rename_func("MAX_BY"), 320 exp.ArgMin: rename_func("MIN_BY"), 321 exp.Array: lambda self, e: f"ARRAY[{self.expressions(e, flat=True)}]", 322 exp.ArrayAny: rename_func("ANY_MATCH"), 323 exp.ArrayConcat: rename_func("CONCAT"), 324 exp.ArrayContains: rename_func("CONTAINS"), 325 exp.ArraySize: rename_func("CARDINALITY"), 326 exp.ArrayUniqueAgg: rename_func("SET_AGG"), 327 exp.AtTimeZone: rename_func("AT_TIMEZONE"), 328 exp.BitwiseAnd: lambda self, e: self.func("BITWISE_AND", e.this, e.expression), 329 exp.BitwiseLeftShift: lambda self, e: self.func( 330 "BITWISE_ARITHMETIC_SHIFT_LEFT", e.this, e.expression 331 ), 332 exp.BitwiseNot: lambda self, e: self.func("BITWISE_NOT", e.this), 333 exp.BitwiseOr: lambda self, e: self.func("BITWISE_OR", e.this, e.expression), 334 exp.BitwiseRightShift: lambda self, e: self.func( 335 "BITWISE_ARITHMETIC_SHIFT_RIGHT", e.this, e.expression 336 ), 337 exp.BitwiseXor: lambda self, e: self.func("BITWISE_XOR", e.this, e.expression), 338 exp.Cast: transforms.preprocess([transforms.epoch_cast_to_ts]), 339 exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP", 340 exp.DateAdd: lambda self, e: self.func( 341 "DATE_ADD", 342 exp.Literal.string(e.text("unit") or "DAY"), 343 _to_int(e.expression), 344 e.this, 345 ), 346 exp.DateDiff: lambda self, e: self.func( 347 "DATE_DIFF", exp.Literal.string(e.text("unit") or "DAY"), e.expression, e.this 348 ), 349 exp.DateStrToDate: datestrtodate_sql, 350 exp.DateToDi: lambda self, 351 e: f"CAST(DATE_FORMAT({self.sql(e, 'this')}, {Presto.DATEINT_FORMAT}) AS INT)", 352 exp.DateSub: lambda self, e: self.func( 353 "DATE_ADD", 354 exp.Literal.string(e.text("unit") or "DAY"), 355 _to_int(e.expression * -1), 356 e.this, 357 ), 358 exp.Decode: lambda self, e: encode_decode_sql(self, e, "FROM_UTF8"), 359 exp.DiToDate: lambda self, 360 e: f"CAST(DATE_PARSE(CAST({self.sql(e, 'this')} AS VARCHAR), {Presto.DATEINT_FORMAT}) AS DATE)", 361 exp.Encode: lambda self, e: encode_decode_sql(self, e, "TO_UTF8"), 362 exp.FileFormatProperty: lambda self, e: f"FORMAT='{e.name.upper()}'", 363 exp.First: _first_last_sql, 364 exp.FirstValue: _first_last_sql, 365 exp.FromTimeZone: lambda self, 366 e: f"WITH_TIMEZONE({self.sql(e, 'this')}, {self.sql(e, 'zone')}) AT TIME ZONE 'UTC'", 367 exp.Group: transforms.preprocess([transforms.unalias_group]), 368 exp.GroupConcat: lambda self, e: self.func( 369 "ARRAY_JOIN", self.func("ARRAY_AGG", e.this), e.args.get("separator") 370 ), 371 exp.Hex: rename_func("TO_HEX"), 372 exp.If: if_sql(), 373 exp.ILike: no_ilike_sql, 374 exp.Initcap: _initcap_sql, 375 exp.ParseJSON: rename_func("JSON_PARSE"), 376 exp.Last: _first_last_sql, 377 exp.LastValue: _first_last_sql, 378 exp.LastDay: lambda self, e: self.func("LAST_DAY_OF_MONTH", e.this), 379 exp.Lateral: _explode_to_unnest_sql, 380 exp.Left: left_to_substring_sql, 381 exp.Levenshtein: rename_func("LEVENSHTEIN_DISTANCE"), 382 exp.LogicalAnd: rename_func("BOOL_AND"), 383 exp.LogicalOr: rename_func("BOOL_OR"), 384 exp.Pivot: no_pivot_sql, 385 exp.Quantile: _quantile_sql, 386 exp.RegexpExtract: regexp_extract_sql, 387 exp.Right: right_to_substring_sql, 388 exp.SafeDivide: no_safe_divide_sql, 389 exp.Schema: _schema_sql, 390 exp.SchemaCommentProperty: lambda self, e: self.naked_property(e), 391 exp.Select: transforms.preprocess( 392 [ 393 transforms.eliminate_qualify, 394 transforms.eliminate_distinct_on, 395 transforms.explode_to_unnest(1), 396 transforms.eliminate_semi_and_anti_joins, 397 ] 398 ), 399 exp.SortArray: _no_sort_array, 400 exp.StrPosition: rename_func("STRPOS"), 401 exp.StrToDate: lambda self, e: f"CAST({_str_to_time_sql(self, e)} AS DATE)", 402 exp.StrToMap: rename_func("SPLIT_TO_MAP"), 403 exp.StrToTime: _str_to_time_sql, 404 exp.StrToUnix: lambda self, e: self.func( 405 "TO_UNIXTIME", self.func("DATE_PARSE", e.this, self.format_time(e)) 406 ), 407 exp.StructExtract: struct_extract_sql, 408 exp.Table: transforms.preprocess([_unnest_sequence]), 409 exp.Timestamp: no_timestamp_sql, 410 exp.TimestampTrunc: timestamptrunc_sql, 411 exp.TimeStrToDate: timestrtotime_sql, 412 exp.TimeStrToTime: timestrtotime_sql, 413 exp.TimeStrToUnix: lambda self, e: self.func( 414 "TO_UNIXTIME", self.func("DATE_PARSE", e.this, Presto.TIME_FORMAT) 415 ), 416 exp.TimeToStr: lambda self, e: self.func("DATE_FORMAT", e.this, self.format_time(e)), 417 exp.TimeToUnix: rename_func("TO_UNIXTIME"), 418 exp.ToChar: lambda self, e: self.func("DATE_FORMAT", e.this, self.format_time(e)), 419 exp.TryCast: transforms.preprocess([transforms.epoch_cast_to_ts]), 420 exp.TsOrDiToDi: lambda self, 421 e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS VARCHAR), '-', ''), 1, 8) AS INT)", 422 exp.TsOrDsAdd: _ts_or_ds_add_sql, 423 exp.TsOrDsDiff: _ts_or_ds_diff_sql, 424 exp.TsOrDsToDate: _ts_or_ds_to_date_sql, 425 exp.Unhex: rename_func("FROM_HEX"), 426 exp.UnixToStr: lambda self, 427 e: f"DATE_FORMAT(FROM_UNIXTIME({self.sql(e, 'this')}), {self.format_time(e)})", 428 exp.UnixToTime: _unix_to_time_sql, 429 exp.UnixToTimeStr: lambda self, 430 e: f"CAST(FROM_UNIXTIME({self.sql(e, 'this')}) AS VARCHAR)", 431 exp.VariancePop: rename_func("VAR_POP"), 432 exp.With: transforms.preprocess([transforms.add_recursive_cte_column_names]), 433 exp.WithinGroup: transforms.preprocess( 434 [transforms.remove_within_group_for_percentiles] 435 ), 436 exp.Xor: bool_xor_sql, 437 } 438 439 def bracket_sql(self, expression: exp.Bracket) -> str: 440 if expression.args.get("safe"): 441 return self.func( 442 "ELEMENT_AT", 443 expression.this, 444 seq_get( 445 apply_index_offset( 446 expression.this, 447 expression.expressions, 448 1 - expression.args.get("offset", 0), 449 ), 450 0, 451 ), 452 ) 453 return super().bracket_sql(expression) 454 455 def struct_sql(self, expression: exp.Struct) -> str: 456 from sqlglot.optimizer.annotate_types import annotate_types 457 458 expression = annotate_types(expression) 459 values: t.List[str] = [] 460 schema: t.List[str] = [] 461 unknown_type = False 462 463 for e in expression.expressions: 464 if isinstance(e, exp.PropertyEQ): 465 if e.type and e.type.is_type(exp.DataType.Type.UNKNOWN): 466 unknown_type = True 467 else: 468 schema.append(f"{self.sql(e, 'this')} {self.sql(e.type)}") 469 values.append(self.sql(e, "expression")) 470 else: 471 values.append(self.sql(e)) 472 473 size = len(expression.expressions) 474 475 if not size or len(schema) != size: 476 if unknown_type: 477 self.unsupported( 478 "Cannot convert untyped key-value definitions (try annotate_types)." 479 ) 480 return self.func("ROW", *values) 481 return f"CAST(ROW({', '.join(values)}) AS ROW({', '.join(schema)}))" 482 483 def interval_sql(self, expression: exp.Interval) -> str: 484 unit = self.sql(expression, "unit") 485 if expression.this and unit.startswith("WEEK"): 486 return f"({expression.this.name} * INTERVAL '7' DAY)" 487 return super().interval_sql(expression) 488 489 def transaction_sql(self, expression: exp.Transaction) -> str: 490 modes = expression.args.get("modes") 491 modes = f" {', '.join(modes)}" if modes else "" 492 return f"START TRANSACTION{modes}" 493 494 def generateseries_sql(self, expression: exp.GenerateSeries) -> str: 495 start = expression.args["start"] 496 end = expression.args["end"] 497 step = expression.args.get("step") 498 499 if isinstance(start, exp.Cast): 500 target_type = start.to 501 elif isinstance(end, exp.Cast): 502 target_type = end.to 503 else: 504 target_type = None 505 506 if target_type and target_type.is_type("timestamp"): 507 if target_type is start.to: 508 end = exp.cast(end, target_type) 509 else: 510 start = exp.cast(start, target_type) 511 512 return self.func("SEQUENCE", start, end, step) 513 514 def offset_limit_modifiers( 515 self, expression: exp.Expression, fetch: bool, limit: t.Optional[exp.Fetch | exp.Limit] 516 ) -> t.List[str]: 517 return [ 518 self.sql(expression, "offset"), 519 self.sql(limit), 520 ] 521 522 def create_sql(self, expression: exp.Create) -> str: 523 """ 524 Presto doesn't support CREATE VIEW with expressions (ex: `CREATE VIEW x (cola)` then `(cola)` is the expression), 525 so we need to remove them 526 """ 527 kind = expression.args["kind"] 528 schema = expression.this 529 if kind == "VIEW" and schema.expressions: 530 expression.this.set("expressions", None) 531 return super().create_sql(expression)
191class Presto(Dialect): 192 INDEX_OFFSET = 1 193 NULL_ORDERING = "nulls_are_last" 194 TIME_FORMAT = MySQL.TIME_FORMAT 195 TIME_MAPPING = MySQL.TIME_MAPPING 196 STRICT_STRING_CONCAT = True 197 SUPPORTS_SEMI_ANTI_JOIN = False 198 TYPED_DIVISION = True 199 TABLESAMPLE_SIZE_IS_PERCENT = True 200 201 # https://github.com/trinodb/trino/issues/17 202 # https://github.com/trinodb/trino/issues/12289 203 # https://github.com/prestodb/presto/issues/2863 204 NORMALIZATION_STRATEGY = NormalizationStrategy.CASE_INSENSITIVE 205 206 class Tokenizer(tokens.Tokenizer): 207 UNICODE_STRINGS = [ 208 (prefix + q, q) 209 for q in t.cast(t.List[str], tokens.Tokenizer.QUOTES) 210 for prefix in ("U&", "u&") 211 ] 212 213 KEYWORDS = { 214 **tokens.Tokenizer.KEYWORDS, 215 "START": TokenType.BEGIN, 216 "MATCH_RECOGNIZE": TokenType.MATCH_RECOGNIZE, 217 "ROW": TokenType.STRUCT, 218 "IPADDRESS": TokenType.IPADDRESS, 219 "IPPREFIX": TokenType.IPPREFIX, 220 } 221 222 class Parser(parser.Parser): 223 VALUES_FOLLOWED_BY_PAREN = False 224 225 FUNCTIONS = { 226 **parser.Parser.FUNCTIONS, 227 "ARBITRARY": exp.AnyValue.from_arg_list, 228 "APPROX_DISTINCT": exp.ApproxDistinct.from_arg_list, 229 "APPROX_PERCENTILE": _build_approx_percentile, 230 "BITWISE_AND": binary_from_function(exp.BitwiseAnd), 231 "BITWISE_NOT": lambda args: exp.BitwiseNot(this=seq_get(args, 0)), 232 "BITWISE_OR": binary_from_function(exp.BitwiseOr), 233 "BITWISE_XOR": binary_from_function(exp.BitwiseXor), 234 "CARDINALITY": exp.ArraySize.from_arg_list, 235 "CONTAINS": exp.ArrayContains.from_arg_list, 236 "DATE_ADD": lambda args: exp.DateAdd( 237 this=seq_get(args, 2), expression=seq_get(args, 1), unit=seq_get(args, 0) 238 ), 239 "DATE_DIFF": lambda args: exp.DateDiff( 240 this=seq_get(args, 2), expression=seq_get(args, 1), unit=seq_get(args, 0) 241 ), 242 "DATE_FORMAT": build_formatted_time(exp.TimeToStr, "presto"), 243 "DATE_PARSE": build_formatted_time(exp.StrToTime, "presto"), 244 "DATE_TRUNC": date_trunc_to_time, 245 "ELEMENT_AT": lambda args: exp.Bracket( 246 this=seq_get(args, 0), expressions=[seq_get(args, 1)], offset=1, safe=True 247 ), 248 "FROM_HEX": exp.Unhex.from_arg_list, 249 "FROM_UNIXTIME": _build_from_unixtime, 250 "FROM_UTF8": lambda args: exp.Decode( 251 this=seq_get(args, 0), replace=seq_get(args, 1), charset=exp.Literal.string("utf-8") 252 ), 253 "NOW": exp.CurrentTimestamp.from_arg_list, 254 "REGEXP_EXTRACT": lambda args: exp.RegexpExtract( 255 this=seq_get(args, 0), expression=seq_get(args, 1), group=seq_get(args, 2) 256 ), 257 "REGEXP_REPLACE": lambda args: exp.RegexpReplace( 258 this=seq_get(args, 0), 259 expression=seq_get(args, 1), 260 replacement=seq_get(args, 2) or exp.Literal.string(""), 261 ), 262 "ROW": exp.Struct.from_arg_list, 263 "SEQUENCE": exp.GenerateSeries.from_arg_list, 264 "SET_AGG": exp.ArrayUniqueAgg.from_arg_list, 265 "SPLIT_TO_MAP": exp.StrToMap.from_arg_list, 266 "STRPOS": lambda args: exp.StrPosition( 267 this=seq_get(args, 0), substr=seq_get(args, 1), instance=seq_get(args, 2) 268 ), 269 "TO_CHAR": _build_to_char, 270 "TO_HEX": exp.Hex.from_arg_list, 271 "TO_UNIXTIME": exp.TimeToUnix.from_arg_list, 272 "TO_UTF8": lambda args: exp.Encode( 273 this=seq_get(args, 0), charset=exp.Literal.string("utf-8") 274 ), 275 } 276 277 FUNCTION_PARSERS = parser.Parser.FUNCTION_PARSERS.copy() 278 FUNCTION_PARSERS.pop("TRIM") 279 280 class Generator(generator.Generator): 281 INTERVAL_ALLOWS_PLURAL_FORM = False 282 JOIN_HINTS = False 283 TABLE_HINTS = False 284 QUERY_HINTS = False 285 IS_BOOL_ALLOWED = False 286 TZ_TO_WITH_TIME_ZONE = True 287 NVL2_SUPPORTED = False 288 STRUCT_DELIMITER = ("(", ")") 289 LIMIT_ONLY_LITERALS = True 290 SUPPORTS_SINGLE_ARG_CONCAT = False 291 LIKE_PROPERTY_INSIDE_SCHEMA = True 292 MULTI_ARG_DISTINCT = False 293 294 PROPERTIES_LOCATION = { 295 **generator.Generator.PROPERTIES_LOCATION, 296 exp.LocationProperty: exp.Properties.Location.UNSUPPORTED, 297 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 298 } 299 300 TYPE_MAPPING = { 301 **generator.Generator.TYPE_MAPPING, 302 exp.DataType.Type.INT: "INTEGER", 303 exp.DataType.Type.FLOAT: "REAL", 304 exp.DataType.Type.BINARY: "VARBINARY", 305 exp.DataType.Type.TEXT: "VARCHAR", 306 exp.DataType.Type.TIMETZ: "TIME", 307 exp.DataType.Type.TIMESTAMPTZ: "TIMESTAMP", 308 exp.DataType.Type.STRUCT: "ROW", 309 exp.DataType.Type.DATETIME: "TIMESTAMP", 310 exp.DataType.Type.DATETIME64: "TIMESTAMP", 311 } 312 313 TRANSFORMS = { 314 **generator.Generator.TRANSFORMS, 315 exp.AnyValue: rename_func("ARBITRARY"), 316 exp.ApproxDistinct: lambda self, e: self.func( 317 "APPROX_DISTINCT", e.this, e.args.get("accuracy") 318 ), 319 exp.ApproxQuantile: rename_func("APPROX_PERCENTILE"), 320 exp.ArgMax: rename_func("MAX_BY"), 321 exp.ArgMin: rename_func("MIN_BY"), 322 exp.Array: lambda self, e: f"ARRAY[{self.expressions(e, flat=True)}]", 323 exp.ArrayAny: rename_func("ANY_MATCH"), 324 exp.ArrayConcat: rename_func("CONCAT"), 325 exp.ArrayContains: rename_func("CONTAINS"), 326 exp.ArraySize: rename_func("CARDINALITY"), 327 exp.ArrayUniqueAgg: rename_func("SET_AGG"), 328 exp.AtTimeZone: rename_func("AT_TIMEZONE"), 329 exp.BitwiseAnd: lambda self, e: self.func("BITWISE_AND", e.this, e.expression), 330 exp.BitwiseLeftShift: lambda self, e: self.func( 331 "BITWISE_ARITHMETIC_SHIFT_LEFT", e.this, e.expression 332 ), 333 exp.BitwiseNot: lambda self, e: self.func("BITWISE_NOT", e.this), 334 exp.BitwiseOr: lambda self, e: self.func("BITWISE_OR", e.this, e.expression), 335 exp.BitwiseRightShift: lambda self, e: self.func( 336 "BITWISE_ARITHMETIC_SHIFT_RIGHT", e.this, e.expression 337 ), 338 exp.BitwiseXor: lambda self, e: self.func("BITWISE_XOR", e.this, e.expression), 339 exp.Cast: transforms.preprocess([transforms.epoch_cast_to_ts]), 340 exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP", 341 exp.DateAdd: lambda self, e: self.func( 342 "DATE_ADD", 343 exp.Literal.string(e.text("unit") or "DAY"), 344 _to_int(e.expression), 345 e.this, 346 ), 347 exp.DateDiff: lambda self, e: self.func( 348 "DATE_DIFF", exp.Literal.string(e.text("unit") or "DAY"), e.expression, e.this 349 ), 350 exp.DateStrToDate: datestrtodate_sql, 351 exp.DateToDi: lambda self, 352 e: f"CAST(DATE_FORMAT({self.sql(e, 'this')}, {Presto.DATEINT_FORMAT}) AS INT)", 353 exp.DateSub: lambda self, e: self.func( 354 "DATE_ADD", 355 exp.Literal.string(e.text("unit") or "DAY"), 356 _to_int(e.expression * -1), 357 e.this, 358 ), 359 exp.Decode: lambda self, e: encode_decode_sql(self, e, "FROM_UTF8"), 360 exp.DiToDate: lambda self, 361 e: f"CAST(DATE_PARSE(CAST({self.sql(e, 'this')} AS VARCHAR), {Presto.DATEINT_FORMAT}) AS DATE)", 362 exp.Encode: lambda self, e: encode_decode_sql(self, e, "TO_UTF8"), 363 exp.FileFormatProperty: lambda self, e: f"FORMAT='{e.name.upper()}'", 364 exp.First: _first_last_sql, 365 exp.FirstValue: _first_last_sql, 366 exp.FromTimeZone: lambda self, 367 e: f"WITH_TIMEZONE({self.sql(e, 'this')}, {self.sql(e, 'zone')}) AT TIME ZONE 'UTC'", 368 exp.Group: transforms.preprocess([transforms.unalias_group]), 369 exp.GroupConcat: lambda self, e: self.func( 370 "ARRAY_JOIN", self.func("ARRAY_AGG", e.this), e.args.get("separator") 371 ), 372 exp.Hex: rename_func("TO_HEX"), 373 exp.If: if_sql(), 374 exp.ILike: no_ilike_sql, 375 exp.Initcap: _initcap_sql, 376 exp.ParseJSON: rename_func("JSON_PARSE"), 377 exp.Last: _first_last_sql, 378 exp.LastValue: _first_last_sql, 379 exp.LastDay: lambda self, e: self.func("LAST_DAY_OF_MONTH", e.this), 380 exp.Lateral: _explode_to_unnest_sql, 381 exp.Left: left_to_substring_sql, 382 exp.Levenshtein: rename_func("LEVENSHTEIN_DISTANCE"), 383 exp.LogicalAnd: rename_func("BOOL_AND"), 384 exp.LogicalOr: rename_func("BOOL_OR"), 385 exp.Pivot: no_pivot_sql, 386 exp.Quantile: _quantile_sql, 387 exp.RegexpExtract: regexp_extract_sql, 388 exp.Right: right_to_substring_sql, 389 exp.SafeDivide: no_safe_divide_sql, 390 exp.Schema: _schema_sql, 391 exp.SchemaCommentProperty: lambda self, e: self.naked_property(e), 392 exp.Select: transforms.preprocess( 393 [ 394 transforms.eliminate_qualify, 395 transforms.eliminate_distinct_on, 396 transforms.explode_to_unnest(1), 397 transforms.eliminate_semi_and_anti_joins, 398 ] 399 ), 400 exp.SortArray: _no_sort_array, 401 exp.StrPosition: rename_func("STRPOS"), 402 exp.StrToDate: lambda self, e: f"CAST({_str_to_time_sql(self, e)} AS DATE)", 403 exp.StrToMap: rename_func("SPLIT_TO_MAP"), 404 exp.StrToTime: _str_to_time_sql, 405 exp.StrToUnix: lambda self, e: self.func( 406 "TO_UNIXTIME", self.func("DATE_PARSE", e.this, self.format_time(e)) 407 ), 408 exp.StructExtract: struct_extract_sql, 409 exp.Table: transforms.preprocess([_unnest_sequence]), 410 exp.Timestamp: no_timestamp_sql, 411 exp.TimestampTrunc: timestamptrunc_sql, 412 exp.TimeStrToDate: timestrtotime_sql, 413 exp.TimeStrToTime: timestrtotime_sql, 414 exp.TimeStrToUnix: lambda self, e: self.func( 415 "TO_UNIXTIME", self.func("DATE_PARSE", e.this, Presto.TIME_FORMAT) 416 ), 417 exp.TimeToStr: lambda self, e: self.func("DATE_FORMAT", e.this, self.format_time(e)), 418 exp.TimeToUnix: rename_func("TO_UNIXTIME"), 419 exp.ToChar: lambda self, e: self.func("DATE_FORMAT", e.this, self.format_time(e)), 420 exp.TryCast: transforms.preprocess([transforms.epoch_cast_to_ts]), 421 exp.TsOrDiToDi: lambda self, 422 e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS VARCHAR), '-', ''), 1, 8) AS INT)", 423 exp.TsOrDsAdd: _ts_or_ds_add_sql, 424 exp.TsOrDsDiff: _ts_or_ds_diff_sql, 425 exp.TsOrDsToDate: _ts_or_ds_to_date_sql, 426 exp.Unhex: rename_func("FROM_HEX"), 427 exp.UnixToStr: lambda self, 428 e: f"DATE_FORMAT(FROM_UNIXTIME({self.sql(e, 'this')}), {self.format_time(e)})", 429 exp.UnixToTime: _unix_to_time_sql, 430 exp.UnixToTimeStr: lambda self, 431 e: f"CAST(FROM_UNIXTIME({self.sql(e, 'this')}) AS VARCHAR)", 432 exp.VariancePop: rename_func("VAR_POP"), 433 exp.With: transforms.preprocess([transforms.add_recursive_cte_column_names]), 434 exp.WithinGroup: transforms.preprocess( 435 [transforms.remove_within_group_for_percentiles] 436 ), 437 exp.Xor: bool_xor_sql, 438 } 439 440 def bracket_sql(self, expression: exp.Bracket) -> str: 441 if expression.args.get("safe"): 442 return self.func( 443 "ELEMENT_AT", 444 expression.this, 445 seq_get( 446 apply_index_offset( 447 expression.this, 448 expression.expressions, 449 1 - expression.args.get("offset", 0), 450 ), 451 0, 452 ), 453 ) 454 return super().bracket_sql(expression) 455 456 def struct_sql(self, expression: exp.Struct) -> str: 457 from sqlglot.optimizer.annotate_types import annotate_types 458 459 expression = annotate_types(expression) 460 values: t.List[str] = [] 461 schema: t.List[str] = [] 462 unknown_type = False 463 464 for e in expression.expressions: 465 if isinstance(e, exp.PropertyEQ): 466 if e.type and e.type.is_type(exp.DataType.Type.UNKNOWN): 467 unknown_type = True 468 else: 469 schema.append(f"{self.sql(e, 'this')} {self.sql(e.type)}") 470 values.append(self.sql(e, "expression")) 471 else: 472 values.append(self.sql(e)) 473 474 size = len(expression.expressions) 475 476 if not size or len(schema) != size: 477 if unknown_type: 478 self.unsupported( 479 "Cannot convert untyped key-value definitions (try annotate_types)." 480 ) 481 return self.func("ROW", *values) 482 return f"CAST(ROW({', '.join(values)}) AS ROW({', '.join(schema)}))" 483 484 def interval_sql(self, expression: exp.Interval) -> str: 485 unit = self.sql(expression, "unit") 486 if expression.this and unit.startswith("WEEK"): 487 return f"({expression.this.name} * INTERVAL '7' DAY)" 488 return super().interval_sql(expression) 489 490 def transaction_sql(self, expression: exp.Transaction) -> str: 491 modes = expression.args.get("modes") 492 modes = f" {', '.join(modes)}" if modes else "" 493 return f"START TRANSACTION{modes}" 494 495 def generateseries_sql(self, expression: exp.GenerateSeries) -> str: 496 start = expression.args["start"] 497 end = expression.args["end"] 498 step = expression.args.get("step") 499 500 if isinstance(start, exp.Cast): 501 target_type = start.to 502 elif isinstance(end, exp.Cast): 503 target_type = end.to 504 else: 505 target_type = None 506 507 if target_type and target_type.is_type("timestamp"): 508 if target_type is start.to: 509 end = exp.cast(end, target_type) 510 else: 511 start = exp.cast(start, target_type) 512 513 return self.func("SEQUENCE", start, end, step) 514 515 def offset_limit_modifiers( 516 self, expression: exp.Expression, fetch: bool, limit: t.Optional[exp.Fetch | exp.Limit] 517 ) -> t.List[str]: 518 return [ 519 self.sql(expression, "offset"), 520 self.sql(limit), 521 ] 522 523 def create_sql(self, expression: exp.Create) -> str: 524 """ 525 Presto doesn't support CREATE VIEW with expressions (ex: `CREATE VIEW x (cola)` then `(cola)` is the expression), 526 so we need to remove them 527 """ 528 kind = expression.args["kind"] 529 schema = expression.this 530 if kind == "VIEW" and schema.expressions: 531 expression.this.set("expressions", None) 532 return super().create_sql(expression)
Default NULL
ordering method to use if not explicitly set.
Possible values: "nulls_are_small"
, "nulls_are_large"
, "nulls_are_last"
Associates this dialect's time formats with their equivalent Python strftime
formats.
Whether the behavior of a / b
depends on the types of a
and b
.
False means a / b
is always float division.
True means a / b
is integer division if both a
and b
are integers.
Specifies the strategy according to which identifiers should be normalized.
Inherited Members
- sqlglot.dialects.dialect.Dialect
- Dialect
- WEEK_OFFSET
- UNNEST_COLUMN_ONLY
- ALIAS_POST_TABLESAMPLE
- IDENTIFIERS_CAN_START_WITH_DIGIT
- DPIPE_IS_STRING_CONCAT
- SUPPORTS_USER_DEFINED_TYPES
- NORMALIZE_FUNCTIONS
- LOG_BASE_FIRST
- SAFE_DIVISION
- CONCAT_COALESCE
- DATE_FORMAT
- DATEINT_FORMAT
- FORMAT_MAPPING
- ESCAPE_SEQUENCES
- PSEUDOCOLUMNS
- PREFER_CTE_ALIAS_COLUMN
- get_or_raise
- format_time
- normalize_identifier
- case_sensitive
- can_identify
- quote_identifier
- to_json_path
- parse
- parse_into
- generate
- transpile
- tokenize
- tokenizer
- parser
- generator
206 class Tokenizer(tokens.Tokenizer): 207 UNICODE_STRINGS = [ 208 (prefix + q, q) 209 for q in t.cast(t.List[str], tokens.Tokenizer.QUOTES) 210 for prefix in ("U&", "u&") 211 ] 212 213 KEYWORDS = { 214 **tokens.Tokenizer.KEYWORDS, 215 "START": TokenType.BEGIN, 216 "MATCH_RECOGNIZE": TokenType.MATCH_RECOGNIZE, 217 "ROW": TokenType.STRUCT, 218 "IPADDRESS": TokenType.IPADDRESS, 219 "IPPREFIX": TokenType.IPPREFIX, 220 }
Inherited Members
- sqlglot.tokens.Tokenizer
- Tokenizer
- SINGLE_TOKENS
- BIT_STRINGS
- BYTE_STRINGS
- HEX_STRINGS
- RAW_STRINGS
- HEREDOC_STRINGS
- IDENTIFIERS
- IDENTIFIER_ESCAPES
- QUOTES
- STRING_ESCAPES
- VAR_SINGLE_TOKENS
- HEREDOC_TAG_IS_IDENTIFIER
- HEREDOC_STRING_ALTERNATIVE
- WHITE_SPACE
- COMMANDS
- COMMAND_PREFIX_TOKENS
- NUMERIC_LITERALS
- COMMENTS
- dialect
- reset
- tokenize
- tokenize_rs
- size
- sql
- tokens
222 class Parser(parser.Parser): 223 VALUES_FOLLOWED_BY_PAREN = False 224 225 FUNCTIONS = { 226 **parser.Parser.FUNCTIONS, 227 "ARBITRARY": exp.AnyValue.from_arg_list, 228 "APPROX_DISTINCT": exp.ApproxDistinct.from_arg_list, 229 "APPROX_PERCENTILE": _build_approx_percentile, 230 "BITWISE_AND": binary_from_function(exp.BitwiseAnd), 231 "BITWISE_NOT": lambda args: exp.BitwiseNot(this=seq_get(args, 0)), 232 "BITWISE_OR": binary_from_function(exp.BitwiseOr), 233 "BITWISE_XOR": binary_from_function(exp.BitwiseXor), 234 "CARDINALITY": exp.ArraySize.from_arg_list, 235 "CONTAINS": exp.ArrayContains.from_arg_list, 236 "DATE_ADD": lambda args: exp.DateAdd( 237 this=seq_get(args, 2), expression=seq_get(args, 1), unit=seq_get(args, 0) 238 ), 239 "DATE_DIFF": lambda args: exp.DateDiff( 240 this=seq_get(args, 2), expression=seq_get(args, 1), unit=seq_get(args, 0) 241 ), 242 "DATE_FORMAT": build_formatted_time(exp.TimeToStr, "presto"), 243 "DATE_PARSE": build_formatted_time(exp.StrToTime, "presto"), 244 "DATE_TRUNC": date_trunc_to_time, 245 "ELEMENT_AT": lambda args: exp.Bracket( 246 this=seq_get(args, 0), expressions=[seq_get(args, 1)], offset=1, safe=True 247 ), 248 "FROM_HEX": exp.Unhex.from_arg_list, 249 "FROM_UNIXTIME": _build_from_unixtime, 250 "FROM_UTF8": lambda args: exp.Decode( 251 this=seq_get(args, 0), replace=seq_get(args, 1), charset=exp.Literal.string("utf-8") 252 ), 253 "NOW": exp.CurrentTimestamp.from_arg_list, 254 "REGEXP_EXTRACT": lambda args: exp.RegexpExtract( 255 this=seq_get(args, 0), expression=seq_get(args, 1), group=seq_get(args, 2) 256 ), 257 "REGEXP_REPLACE": lambda args: exp.RegexpReplace( 258 this=seq_get(args, 0), 259 expression=seq_get(args, 1), 260 replacement=seq_get(args, 2) or exp.Literal.string(""), 261 ), 262 "ROW": exp.Struct.from_arg_list, 263 "SEQUENCE": exp.GenerateSeries.from_arg_list, 264 "SET_AGG": exp.ArrayUniqueAgg.from_arg_list, 265 "SPLIT_TO_MAP": exp.StrToMap.from_arg_list, 266 "STRPOS": lambda args: exp.StrPosition( 267 this=seq_get(args, 0), substr=seq_get(args, 1), instance=seq_get(args, 2) 268 ), 269 "TO_CHAR": _build_to_char, 270 "TO_HEX": exp.Hex.from_arg_list, 271 "TO_UNIXTIME": exp.TimeToUnix.from_arg_list, 272 "TO_UTF8": lambda args: exp.Encode( 273 this=seq_get(args, 0), charset=exp.Literal.string("utf-8") 274 ), 275 } 276 277 FUNCTION_PARSERS = parser.Parser.FUNCTION_PARSERS.copy() 278 FUNCTION_PARSERS.pop("TRIM")
Parser consumes a list of tokens produced by the Tokenizer and produces a parsed syntax tree.
Arguments:
- error_level: The desired error level. Default: ErrorLevel.IMMEDIATE
- error_message_context: The amount of context to capture from a query string when displaying the error message (in number of characters). Default: 100
- max_errors: Maximum number of error messages to include in a raised ParseError. This is only relevant if error_level is ErrorLevel.RAISE. Default: 3
Inherited Members
- sqlglot.parser.Parser
- Parser
- NO_PAREN_FUNCTIONS
- STRUCT_TYPE_TOKENS
- NESTED_TYPE_TOKENS
- ENUM_TYPE_TOKENS
- AGGREGATE_TYPE_TOKENS
- TYPE_TOKENS
- SIGNED_TO_UNSIGNED_TYPE_TOKEN
- SUBQUERY_PREDICATES
- RESERVED_TOKENS
- DB_CREATABLES
- CREATABLES
- ID_VAR_TOKENS
- INTERVAL_VARS
- COMMENT_TABLE_ALIAS_TOKENS
- UPDATE_ALIAS_TOKENS
- TRIM_TYPES
- FUNC_TOKENS
- CONJUNCTION
- EQUALITY
- COMPARISON
- BITWISE
- TERM
- FACTOR
- EXPONENT
- TIMES
- TIMESTAMPS
- SET_OPERATIONS
- JOIN_METHODS
- JOIN_SIDES
- JOIN_KINDS
- JOIN_HINTS
- LAMBDAS
- COLUMN_OPERATORS
- EXPRESSION_PARSERS
- STATEMENT_PARSERS
- UNARY_PARSERS
- PRIMARY_PARSERS
- PLACEHOLDER_PARSERS
- RANGE_PARSERS
- PROPERTY_PARSERS
- CONSTRAINT_PARSERS
- ALTER_PARSERS
- SCHEMA_UNNAMED_CONSTRAINTS
- NO_PAREN_FUNCTION_PARSERS
- INVALID_FUNC_NAME_TOKENS
- FUNCTIONS_WITH_ALIASED_ARGS
- KEY_VALUE_DEFINITIONS
- QUERY_MODIFIER_PARSERS
- SET_PARSERS
- SHOW_PARSERS
- TYPE_LITERAL_PARSERS
- DDL_SELECT_TOKENS
- PRE_VOLATILE_TOKENS
- TRANSACTION_KIND
- TRANSACTION_CHARACTERISTICS
- USABLES
- INSERT_ALTERNATIVES
- CLONE_KEYWORDS
- HISTORICAL_DATA_KIND
- OPCLASS_FOLLOW_KEYWORDS
- OPTYPE_FOLLOW_TOKENS
- TABLE_INDEX_HINT_TOKENS
- WINDOW_ALIAS_TOKENS
- WINDOW_BEFORE_PAREN_TOKENS
- WINDOW_SIDES
- JSON_KEY_VALUE_SEPARATOR_TOKENS
- FETCH_TOKENS
- ADD_CONSTRAINT_TOKENS
- DISTINCT_TOKENS
- NULL_TOKENS
- UNNEST_OFFSET_ALIAS_TOKENS
- STRICT_CAST
- PREFIXED_PIVOT_COLUMNS
- IDENTIFY_PIVOT_STRINGS
- LOG_DEFAULTS_TO_LN
- ALTER_TABLE_ADD_REQUIRED_FOR_EACH_COLUMN
- TABLESAMPLE_CSV
- SET_REQUIRES_ASSIGNMENT_DELIMITER
- TRIM_PATTERN_FIRST
- STRING_ALIASES
- MODIFIERS_ATTACHED_TO_UNION
- UNION_MODIFIERS
- NO_PAREN_IF_COMMANDS
- JSON_ARROWS_REQUIRE_JSON_TYPE
- SUPPORTS_IMPLICIT_UNNEST
- error_level
- error_message_context
- max_errors
- dialect
- reset
- parse
- parse_into
- check_errors
- raise_error
- expression
- validate_expression
- errors
- sql
280 class Generator(generator.Generator): 281 INTERVAL_ALLOWS_PLURAL_FORM = False 282 JOIN_HINTS = False 283 TABLE_HINTS = False 284 QUERY_HINTS = False 285 IS_BOOL_ALLOWED = False 286 TZ_TO_WITH_TIME_ZONE = True 287 NVL2_SUPPORTED = False 288 STRUCT_DELIMITER = ("(", ")") 289 LIMIT_ONLY_LITERALS = True 290 SUPPORTS_SINGLE_ARG_CONCAT = False 291 LIKE_PROPERTY_INSIDE_SCHEMA = True 292 MULTI_ARG_DISTINCT = False 293 294 PROPERTIES_LOCATION = { 295 **generator.Generator.PROPERTIES_LOCATION, 296 exp.LocationProperty: exp.Properties.Location.UNSUPPORTED, 297 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 298 } 299 300 TYPE_MAPPING = { 301 **generator.Generator.TYPE_MAPPING, 302 exp.DataType.Type.INT: "INTEGER", 303 exp.DataType.Type.FLOAT: "REAL", 304 exp.DataType.Type.BINARY: "VARBINARY", 305 exp.DataType.Type.TEXT: "VARCHAR", 306 exp.DataType.Type.TIMETZ: "TIME", 307 exp.DataType.Type.TIMESTAMPTZ: "TIMESTAMP", 308 exp.DataType.Type.STRUCT: "ROW", 309 exp.DataType.Type.DATETIME: "TIMESTAMP", 310 exp.DataType.Type.DATETIME64: "TIMESTAMP", 311 } 312 313 TRANSFORMS = { 314 **generator.Generator.TRANSFORMS, 315 exp.AnyValue: rename_func("ARBITRARY"), 316 exp.ApproxDistinct: lambda self, e: self.func( 317 "APPROX_DISTINCT", e.this, e.args.get("accuracy") 318 ), 319 exp.ApproxQuantile: rename_func("APPROX_PERCENTILE"), 320 exp.ArgMax: rename_func("MAX_BY"), 321 exp.ArgMin: rename_func("MIN_BY"), 322 exp.Array: lambda self, e: f"ARRAY[{self.expressions(e, flat=True)}]", 323 exp.ArrayAny: rename_func("ANY_MATCH"), 324 exp.ArrayConcat: rename_func("CONCAT"), 325 exp.ArrayContains: rename_func("CONTAINS"), 326 exp.ArraySize: rename_func("CARDINALITY"), 327 exp.ArrayUniqueAgg: rename_func("SET_AGG"), 328 exp.AtTimeZone: rename_func("AT_TIMEZONE"), 329 exp.BitwiseAnd: lambda self, e: self.func("BITWISE_AND", e.this, e.expression), 330 exp.BitwiseLeftShift: lambda self, e: self.func( 331 "BITWISE_ARITHMETIC_SHIFT_LEFT", e.this, e.expression 332 ), 333 exp.BitwiseNot: lambda self, e: self.func("BITWISE_NOT", e.this), 334 exp.BitwiseOr: lambda self, e: self.func("BITWISE_OR", e.this, e.expression), 335 exp.BitwiseRightShift: lambda self, e: self.func( 336 "BITWISE_ARITHMETIC_SHIFT_RIGHT", e.this, e.expression 337 ), 338 exp.BitwiseXor: lambda self, e: self.func("BITWISE_XOR", e.this, e.expression), 339 exp.Cast: transforms.preprocess([transforms.epoch_cast_to_ts]), 340 exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP", 341 exp.DateAdd: lambda self, e: self.func( 342 "DATE_ADD", 343 exp.Literal.string(e.text("unit") or "DAY"), 344 _to_int(e.expression), 345 e.this, 346 ), 347 exp.DateDiff: lambda self, e: self.func( 348 "DATE_DIFF", exp.Literal.string(e.text("unit") or "DAY"), e.expression, e.this 349 ), 350 exp.DateStrToDate: datestrtodate_sql, 351 exp.DateToDi: lambda self, 352 e: f"CAST(DATE_FORMAT({self.sql(e, 'this')}, {Presto.DATEINT_FORMAT}) AS INT)", 353 exp.DateSub: lambda self, e: self.func( 354 "DATE_ADD", 355 exp.Literal.string(e.text("unit") or "DAY"), 356 _to_int(e.expression * -1), 357 e.this, 358 ), 359 exp.Decode: lambda self, e: encode_decode_sql(self, e, "FROM_UTF8"), 360 exp.DiToDate: lambda self, 361 e: f"CAST(DATE_PARSE(CAST({self.sql(e, 'this')} AS VARCHAR), {Presto.DATEINT_FORMAT}) AS DATE)", 362 exp.Encode: lambda self, e: encode_decode_sql(self, e, "TO_UTF8"), 363 exp.FileFormatProperty: lambda self, e: f"FORMAT='{e.name.upper()}'", 364 exp.First: _first_last_sql, 365 exp.FirstValue: _first_last_sql, 366 exp.FromTimeZone: lambda self, 367 e: f"WITH_TIMEZONE({self.sql(e, 'this')}, {self.sql(e, 'zone')}) AT TIME ZONE 'UTC'", 368 exp.Group: transforms.preprocess([transforms.unalias_group]), 369 exp.GroupConcat: lambda self, e: self.func( 370 "ARRAY_JOIN", self.func("ARRAY_AGG", e.this), e.args.get("separator") 371 ), 372 exp.Hex: rename_func("TO_HEX"), 373 exp.If: if_sql(), 374 exp.ILike: no_ilike_sql, 375 exp.Initcap: _initcap_sql, 376 exp.ParseJSON: rename_func("JSON_PARSE"), 377 exp.Last: _first_last_sql, 378 exp.LastValue: _first_last_sql, 379 exp.LastDay: lambda self, e: self.func("LAST_DAY_OF_MONTH", e.this), 380 exp.Lateral: _explode_to_unnest_sql, 381 exp.Left: left_to_substring_sql, 382 exp.Levenshtein: rename_func("LEVENSHTEIN_DISTANCE"), 383 exp.LogicalAnd: rename_func("BOOL_AND"), 384 exp.LogicalOr: rename_func("BOOL_OR"), 385 exp.Pivot: no_pivot_sql, 386 exp.Quantile: _quantile_sql, 387 exp.RegexpExtract: regexp_extract_sql, 388 exp.Right: right_to_substring_sql, 389 exp.SafeDivide: no_safe_divide_sql, 390 exp.Schema: _schema_sql, 391 exp.SchemaCommentProperty: lambda self, e: self.naked_property(e), 392 exp.Select: transforms.preprocess( 393 [ 394 transforms.eliminate_qualify, 395 transforms.eliminate_distinct_on, 396 transforms.explode_to_unnest(1), 397 transforms.eliminate_semi_and_anti_joins, 398 ] 399 ), 400 exp.SortArray: _no_sort_array, 401 exp.StrPosition: rename_func("STRPOS"), 402 exp.StrToDate: lambda self, e: f"CAST({_str_to_time_sql(self, e)} AS DATE)", 403 exp.StrToMap: rename_func("SPLIT_TO_MAP"), 404 exp.StrToTime: _str_to_time_sql, 405 exp.StrToUnix: lambda self, e: self.func( 406 "TO_UNIXTIME", self.func("DATE_PARSE", e.this, self.format_time(e)) 407 ), 408 exp.StructExtract: struct_extract_sql, 409 exp.Table: transforms.preprocess([_unnest_sequence]), 410 exp.Timestamp: no_timestamp_sql, 411 exp.TimestampTrunc: timestamptrunc_sql, 412 exp.TimeStrToDate: timestrtotime_sql, 413 exp.TimeStrToTime: timestrtotime_sql, 414 exp.TimeStrToUnix: lambda self, e: self.func( 415 "TO_UNIXTIME", self.func("DATE_PARSE", e.this, Presto.TIME_FORMAT) 416 ), 417 exp.TimeToStr: lambda self, e: self.func("DATE_FORMAT", e.this, self.format_time(e)), 418 exp.TimeToUnix: rename_func("TO_UNIXTIME"), 419 exp.ToChar: lambda self, e: self.func("DATE_FORMAT", e.this, self.format_time(e)), 420 exp.TryCast: transforms.preprocess([transforms.epoch_cast_to_ts]), 421 exp.TsOrDiToDi: lambda self, 422 e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS VARCHAR), '-', ''), 1, 8) AS INT)", 423 exp.TsOrDsAdd: _ts_or_ds_add_sql, 424 exp.TsOrDsDiff: _ts_or_ds_diff_sql, 425 exp.TsOrDsToDate: _ts_or_ds_to_date_sql, 426 exp.Unhex: rename_func("FROM_HEX"), 427 exp.UnixToStr: lambda self, 428 e: f"DATE_FORMAT(FROM_UNIXTIME({self.sql(e, 'this')}), {self.format_time(e)})", 429 exp.UnixToTime: _unix_to_time_sql, 430 exp.UnixToTimeStr: lambda self, 431 e: f"CAST(FROM_UNIXTIME({self.sql(e, 'this')}) AS VARCHAR)", 432 exp.VariancePop: rename_func("VAR_POP"), 433 exp.With: transforms.preprocess([transforms.add_recursive_cte_column_names]), 434 exp.WithinGroup: transforms.preprocess( 435 [transforms.remove_within_group_for_percentiles] 436 ), 437 exp.Xor: bool_xor_sql, 438 } 439 440 def bracket_sql(self, expression: exp.Bracket) -> str: 441 if expression.args.get("safe"): 442 return self.func( 443 "ELEMENT_AT", 444 expression.this, 445 seq_get( 446 apply_index_offset( 447 expression.this, 448 expression.expressions, 449 1 - expression.args.get("offset", 0), 450 ), 451 0, 452 ), 453 ) 454 return super().bracket_sql(expression) 455 456 def struct_sql(self, expression: exp.Struct) -> str: 457 from sqlglot.optimizer.annotate_types import annotate_types 458 459 expression = annotate_types(expression) 460 values: t.List[str] = [] 461 schema: t.List[str] = [] 462 unknown_type = False 463 464 for e in expression.expressions: 465 if isinstance(e, exp.PropertyEQ): 466 if e.type and e.type.is_type(exp.DataType.Type.UNKNOWN): 467 unknown_type = True 468 else: 469 schema.append(f"{self.sql(e, 'this')} {self.sql(e.type)}") 470 values.append(self.sql(e, "expression")) 471 else: 472 values.append(self.sql(e)) 473 474 size = len(expression.expressions) 475 476 if not size or len(schema) != size: 477 if unknown_type: 478 self.unsupported( 479 "Cannot convert untyped key-value definitions (try annotate_types)." 480 ) 481 return self.func("ROW", *values) 482 return f"CAST(ROW({', '.join(values)}) AS ROW({', '.join(schema)}))" 483 484 def interval_sql(self, expression: exp.Interval) -> str: 485 unit = self.sql(expression, "unit") 486 if expression.this and unit.startswith("WEEK"): 487 return f"({expression.this.name} * INTERVAL '7' DAY)" 488 return super().interval_sql(expression) 489 490 def transaction_sql(self, expression: exp.Transaction) -> str: 491 modes = expression.args.get("modes") 492 modes = f" {', '.join(modes)}" if modes else "" 493 return f"START TRANSACTION{modes}" 494 495 def generateseries_sql(self, expression: exp.GenerateSeries) -> str: 496 start = expression.args["start"] 497 end = expression.args["end"] 498 step = expression.args.get("step") 499 500 if isinstance(start, exp.Cast): 501 target_type = start.to 502 elif isinstance(end, exp.Cast): 503 target_type = end.to 504 else: 505 target_type = None 506 507 if target_type and target_type.is_type("timestamp"): 508 if target_type is start.to: 509 end = exp.cast(end, target_type) 510 else: 511 start = exp.cast(start, target_type) 512 513 return self.func("SEQUENCE", start, end, step) 514 515 def offset_limit_modifiers( 516 self, expression: exp.Expression, fetch: bool, limit: t.Optional[exp.Fetch | exp.Limit] 517 ) -> t.List[str]: 518 return [ 519 self.sql(expression, "offset"), 520 self.sql(limit), 521 ] 522 523 def create_sql(self, expression: exp.Create) -> str: 524 """ 525 Presto doesn't support CREATE VIEW with expressions (ex: `CREATE VIEW x (cola)` then `(cola)` is the expression), 526 so we need to remove them 527 """ 528 kind = expression.args["kind"] 529 schema = expression.this 530 if kind == "VIEW" and schema.expressions: 531 expression.this.set("expressions", None) 532 return super().create_sql(expression)
Generator converts a given syntax tree to the corresponding SQL string.
Arguments:
- pretty: Whether 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 to normalize identifiers to lowercase. Default: False.
- pad: The pad size in a formatted string. Default: 2.
- indent: The indentation size in a formatted string. Default: 2.
- normalize_functions: How to normalize 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: Whether 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 to preserve comments in the output SQL code. Default: True
440 def bracket_sql(self, expression: exp.Bracket) -> str: 441 if expression.args.get("safe"): 442 return self.func( 443 "ELEMENT_AT", 444 expression.this, 445 seq_get( 446 apply_index_offset( 447 expression.this, 448 expression.expressions, 449 1 - expression.args.get("offset", 0), 450 ), 451 0, 452 ), 453 ) 454 return super().bracket_sql(expression)
456 def struct_sql(self, expression: exp.Struct) -> str: 457 from sqlglot.optimizer.annotate_types import annotate_types 458 459 expression = annotate_types(expression) 460 values: t.List[str] = [] 461 schema: t.List[str] = [] 462 unknown_type = False 463 464 for e in expression.expressions: 465 if isinstance(e, exp.PropertyEQ): 466 if e.type and e.type.is_type(exp.DataType.Type.UNKNOWN): 467 unknown_type = True 468 else: 469 schema.append(f"{self.sql(e, 'this')} {self.sql(e.type)}") 470 values.append(self.sql(e, "expression")) 471 else: 472 values.append(self.sql(e)) 473 474 size = len(expression.expressions) 475 476 if not size or len(schema) != size: 477 if unknown_type: 478 self.unsupported( 479 "Cannot convert untyped key-value definitions (try annotate_types)." 480 ) 481 return self.func("ROW", *values) 482 return f"CAST(ROW({', '.join(values)}) AS ROW({', '.join(schema)}))"
495 def generateseries_sql(self, expression: exp.GenerateSeries) -> str: 496 start = expression.args["start"] 497 end = expression.args["end"] 498 step = expression.args.get("step") 499 500 if isinstance(start, exp.Cast): 501 target_type = start.to 502 elif isinstance(end, exp.Cast): 503 target_type = end.to 504 else: 505 target_type = None 506 507 if target_type and target_type.is_type("timestamp"): 508 if target_type is start.to: 509 end = exp.cast(end, target_type) 510 else: 511 start = exp.cast(start, target_type) 512 513 return self.func("SEQUENCE", start, end, step)
523 def create_sql(self, expression: exp.Create) -> str: 524 """ 525 Presto doesn't support CREATE VIEW with expressions (ex: `CREATE VIEW x (cola)` then `(cola)` is the expression), 526 so we need to remove them 527 """ 528 kind = expression.args["kind"] 529 schema = expression.this 530 if kind == "VIEW" and schema.expressions: 531 expression.this.set("expressions", None) 532 return super().create_sql(expression)
Presto doesn't support CREATE VIEW with expressions (ex: CREATE VIEW x (cola)
then (cola)
is the expression),
so we need to remove them
Inherited Members
- sqlglot.generator.Generator
- Generator
- NULL_ORDERING_SUPPORTED
- IGNORE_NULLS_IN_FUNC
- LOCKING_READS_SUPPORTED
- EXPLICIT_UNION
- WRAP_DERIVED_VALUES
- CREATE_FUNCTION_RETURN_AS
- MATCHED_BY_SOURCE
- SINGLE_STRING_INTERVAL
- LIMIT_FETCH
- RENAME_TABLE_WITH_DB
- GROUPINGS_SEP
- INDEX_ON
- QUERY_HINT_SEP
- DUPLICATE_KEY_UPDATE_WITH_SET
- LIMIT_IS_TOP
- RETURNING_END
- COLUMN_JOIN_MARKS_SUPPORTED
- EXTRACT_ALLOWS_QUOTES
- VALUES_AS_TABLE
- ALTER_TABLE_INCLUDE_COLUMN_KEYWORD
- UNNEST_WITH_ORDINALITY
- AGGREGATE_FILTER_SUPPORTED
- SEMI_ANTI_JOIN_WITH_SIDE
- COMPUTED_COLUMN_WITH_TYPE
- SUPPORTS_TABLE_COPY
- TABLESAMPLE_REQUIRES_PARENS
- TABLESAMPLE_SIZE_IS_ROWS
- TABLESAMPLE_KEYWORDS
- TABLESAMPLE_WITH_METHOD
- TABLESAMPLE_SEED_KEYWORD
- COLLATE_IS_FUNC
- DATA_TYPE_SPECIFIERS_ALLOWED
- ENSURE_BOOLS
- CTE_RECURSIVE_KEYWORD_REQUIRED
- LAST_DAY_SUPPORTS_DATE_PART
- SUPPORTS_TABLE_ALIAS_COLUMNS
- UNPIVOT_ALIASES_ARE_IDENTIFIERS
- JSON_KEY_VALUE_PAIR_SEP
- INSERT_OVERWRITE
- SUPPORTS_SELECT_INTO
- SUPPORTS_UNLOGGED_TABLES
- SUPPORTS_CREATE_TABLE_LIKE
- JSON_TYPE_REQUIRED_FOR_EXTRACTION
- JSON_PATH_BRACKETED_KEY_SUPPORTED
- JSON_PATH_SINGLE_QUOTE_ESCAPE
- SUPPORTED_JSON_PATH_PARTS
- CAN_IMPLEMENT_ARRAY_ANY
- STAR_MAPPING
- TIME_PART_SINGULARS
- TOKEN_MAPPING
- PARAMETER_TOKEN
- NAMED_PLACEHOLDER_TOKEN
- RESERVED_KEYWORDS
- WITH_SEPARATED_COMMENTS
- EXCLUDE_COMMENTS
- UNWRAPPED_INTERVAL_VALUES
- PARAMETERIZABLE_TEXT_TYPES
- EXPRESSIONS_WITHOUT_NESTED_CTES
- SENTINEL_LINE_BREAK
- pretty
- identify
- normalize
- pad
- unsupported_level
- max_unsupported
- leading_comma
- max_text_width
- comments
- dialect
- normalize_functions
- unsupported_messages
- generate
- preprocess
- unsupported
- sep
- seg
- pad_comment
- maybe_comment
- wrap
- no_identify
- normalize_func
- indent
- sql
- uncache_sql
- cache_sql
- characterset_sql
- column_sql
- columnposition_sql
- columndef_sql
- columnconstraint_sql
- computedcolumnconstraint_sql
- autoincrementcolumnconstraint_sql
- compresscolumnconstraint_sql
- generatedasidentitycolumnconstraint_sql
- generatedasrowcolumnconstraint_sql
- periodforsystemtimeconstraint_sql
- notnullcolumnconstraint_sql
- transformcolumnconstraint_sql
- primarykeycolumnconstraint_sql
- uniquecolumnconstraint_sql
- createable_sql
- clone_sql
- describe_sql
- heredoc_sql
- prepend_ctes
- with_sql
- cte_sql
- tablealias_sql
- bitstring_sql
- hexstring_sql
- bytestring_sql
- unicodestring_sql
- rawstring_sql
- datatypeparam_sql
- datatype_sql
- directory_sql
- delete_sql
- drop_sql
- except_sql
- except_op
- fetch_sql
- filter_sql
- hint_sql
- index_sql
- identifier_sql
- inputoutputformat_sql
- national_sql
- partition_sql
- properties_sql
- root_properties
- properties
- with_properties
- locate_properties
- property_name
- property_sql
- likeproperty_sql
- fallbackproperty_sql
- journalproperty_sql
- freespaceproperty_sql
- checksumproperty_sql
- mergeblockratioproperty_sql
- datablocksizeproperty_sql
- blockcompressionproperty_sql
- isolatedloadingproperty_sql
- partitionboundspec_sql
- partitionedofproperty_sql
- lockingproperty_sql
- withdataproperty_sql
- withsystemversioningproperty_sql
- insert_sql
- intersect_sql
- intersect_op
- introducer_sql
- kill_sql
- pseudotype_sql
- objectidentifier_sql
- onconflict_sql
- returning_sql
- rowformatdelimitedproperty_sql
- withtablehint_sql
- indextablehint_sql
- historicaldata_sql
- table_parts
- table_sql
- tablesample_sql
- pivot_sql
- version_sql
- tuple_sql
- update_sql
- values_sql
- var_sql
- into_sql
- from_sql
- group_sql
- having_sql
- connect_sql
- prior_sql
- join_sql
- lambda_sql
- lateral_op
- lateral_sql
- limit_sql
- offset_sql
- setitem_sql
- set_sql
- pragma_sql
- lock_sql
- literal_sql
- escape_str
- loaddata_sql
- null_sql
- boolean_sql
- order_sql
- withfill_sql
- cluster_sql
- distribute_sql
- sort_sql
- ordered_sql
- matchrecognize_sql
- query_modifiers
- queryoption_sql
- after_having_modifiers
- after_limit_modifiers
- select_sql
- schema_sql
- schema_columns_sql
- star_sql
- parameter_sql
- sessionparameter_sql
- placeholder_sql
- subquery_sql
- qualify_sql
- union_sql
- union_op
- unnest_sql
- prewhere_sql
- where_sql
- window_sql
- partition_by_sql
- windowspec_sql
- withingroup_sql
- between_sql
- all_sql
- any_sql
- exists_sql
- case_sql
- constraint_sql
- nextvaluefor_sql
- extract_sql
- trim_sql
- convert_concat_args
- concat_sql
- concatws_sql
- check_sql
- foreignkey_sql
- primarykey_sql
- if_sql
- matchagainst_sql
- jsonkeyvalue_sql
- jsonpath_sql
- json_path_part
- formatjson_sql
- jsonobject_sql
- jsonobjectagg_sql
- jsonarray_sql
- jsonarrayagg_sql
- jsoncolumndef_sql
- jsonschema_sql
- jsontable_sql
- openjsoncolumndef_sql
- openjson_sql
- in_sql
- in_unnest_op
- return_sql
- reference_sql
- anonymous_sql
- paren_sql
- neg_sql
- not_sql
- alias_sql
- pivotalias_sql
- aliases_sql
- atindex_sql
- attimezone_sql
- fromtimezone_sql
- add_sql
- and_sql
- xor_sql
- connector_sql
- bitwiseand_sql
- bitwiseleftshift_sql
- bitwisenot_sql
- bitwiseor_sql
- bitwiserightshift_sql
- bitwisexor_sql
- cast_sql
- currentdate_sql
- currenttimestamp_sql
- collate_sql
- command_sql
- comment_sql
- mergetreettlaction_sql
- mergetreettl_sql
- commit_sql
- rollback_sql
- altercolumn_sql
- renametable_sql
- renamecolumn_sql
- altertable_sql
- add_column_sql
- droppartition_sql
- addconstraint_sql
- distinct_sql
- ignorenulls_sql
- respectnulls_sql
- havingmax_sql
- intdiv_sql
- dpipe_sql
- div_sql
- overlaps_sql
- distance_sql
- dot_sql
- eq_sql
- propertyeq_sql
- escape_sql
- glob_sql
- gt_sql
- gte_sql
- ilike_sql
- ilikeany_sql
- is_sql
- like_sql
- likeany_sql
- similarto_sql
- lt_sql
- lte_sql
- mod_sql
- mul_sql
- neq_sql
- nullsafeeq_sql
- nullsafeneq_sql
- or_sql
- slice_sql
- sub_sql
- trycast_sql
- log_sql
- use_sql
- binary
- function_fallback_sql
- func
- format_args
- text_width
- format_time
- expressions
- op_expressions
- naked_property
- set_operation
- tag_sql
- token_sql
- userdefinedfunction_sql
- joinhint_sql
- kwarg_sql
- when_sql
- merge_sql
- tochar_sql
- dictproperty_sql
- dictrange_sql
- dictsubproperty_sql
- oncluster_sql
- clusteredbyproperty_sql
- anyvalue_sql
- querytransform_sql
- indexconstraintoption_sql
- checkcolumnconstraint_sql
- indexcolumnconstraint_sql
- nvl2_sql
- comprehension_sql
- columnprefix_sql
- opclass_sql
- predict_sql
- forin_sql
- refresh_sql
- operator_sql
- toarray_sql
- tsordstotime_sql
- tsordstodate_sql
- unixdate_sql
- lastday_sql
- arrayany_sql
- partitionrange_sql
- truncatetable_sql
- convert_sql