Edit on GitHub

sqlglot.dialects.clickhouse

  1from __future__ import annotations
  2
  3import typing as t
  4
  5from sqlglot import exp, generator, parser, tokens
  6from sqlglot.dialects.dialect import (
  7    Dialect,
  8    inline_array_sql,
  9    no_pivot_sql,
 10    rename_func,
 11    var_map_sql,
 12)
 13from sqlglot.errors import ParseError
 14from sqlglot.parser import parse_var_map
 15from sqlglot.tokens import Token, TokenType
 16
 17
 18def _lower_func(sql: str) -> str:
 19    index = sql.index("(")
 20    return sql[:index].lower() + sql[index:]
 21
 22
 23class ClickHouse(Dialect):
 24    NORMALIZE_FUNCTIONS: bool | str = False
 25    NULL_ORDERING = "nulls_are_last"
 26    STRICT_STRING_CONCAT = True
 27
 28    class Tokenizer(tokens.Tokenizer):
 29        COMMENTS = ["--", "#", "#!", ("/*", "*/")]
 30        IDENTIFIERS = ['"', "`"]
 31        STRING_ESCAPES = ["'", "\\"]
 32        BIT_STRINGS = [("0b", "")]
 33        HEX_STRINGS = [("0x", ""), ("0X", "")]
 34
 35        KEYWORDS = {
 36            **tokens.Tokenizer.KEYWORDS,
 37            "ATTACH": TokenType.COMMAND,
 38            "DATETIME64": TokenType.DATETIME64,
 39            "DICTIONARY": TokenType.DICTIONARY,
 40            "FINAL": TokenType.FINAL,
 41            "FLOAT32": TokenType.FLOAT,
 42            "FLOAT64": TokenType.DOUBLE,
 43            "GLOBAL": TokenType.GLOBAL,
 44            "INT128": TokenType.INT128,
 45            "INT16": TokenType.SMALLINT,
 46            "INT256": TokenType.INT256,
 47            "INT32": TokenType.INT,
 48            "INT64": TokenType.BIGINT,
 49            "INT8": TokenType.TINYINT,
 50            "MAP": TokenType.MAP,
 51            "TUPLE": TokenType.STRUCT,
 52            "UINT128": TokenType.UINT128,
 53            "UINT16": TokenType.USMALLINT,
 54            "UINT256": TokenType.UINT256,
 55            "UINT32": TokenType.UINT,
 56            "UINT64": TokenType.UBIGINT,
 57            "UINT8": TokenType.UTINYINT,
 58        }
 59
 60    class Parser(parser.Parser):
 61        FUNCTIONS = {
 62            **parser.Parser.FUNCTIONS,
 63            "ANY": exp.AnyValue.from_arg_list,
 64            "MAP": parse_var_map,
 65            "MATCH": exp.RegexpLike.from_arg_list,
 66            "UNIQ": exp.ApproxDistinct.from_arg_list,
 67        }
 68
 69        FUNCTIONS_WITH_ALIASED_ARGS = {*parser.Parser.FUNCTIONS_WITH_ALIASED_ARGS, "TUPLE"}
 70
 71        FUNCTION_PARSERS = {
 72            **parser.Parser.FUNCTION_PARSERS,
 73            "QUANTILE": lambda self: self._parse_quantile(),
 74        }
 75
 76        FUNCTION_PARSERS.pop("MATCH")
 77
 78        NO_PAREN_FUNCTION_PARSERS = parser.Parser.NO_PAREN_FUNCTION_PARSERS.copy()
 79        NO_PAREN_FUNCTION_PARSERS.pop(TokenType.ANY)
 80
 81        RANGE_PARSERS = {
 82            **parser.Parser.RANGE_PARSERS,
 83            TokenType.GLOBAL: lambda self, this: self._match(TokenType.IN)
 84            and self._parse_in(this, is_global=True),
 85        }
 86
 87        # The PLACEHOLDER entry is popped because 1) it doesn't affect Clickhouse (it corresponds to
 88        # the postgres-specific JSONBContains parser) and 2) it makes parsing the ternary op simpler.
 89        COLUMN_OPERATORS = parser.Parser.COLUMN_OPERATORS.copy()
 90        COLUMN_OPERATORS.pop(TokenType.PLACEHOLDER)
 91
 92        JOIN_KINDS = {
 93            *parser.Parser.JOIN_KINDS,
 94            TokenType.ANY,
 95            TokenType.ASOF,
 96            TokenType.ANTI,
 97            TokenType.SEMI,
 98        }
 99
100        TABLE_ALIAS_TOKENS = {*parser.Parser.TABLE_ALIAS_TOKENS} - {
101            TokenType.ANY,
102            TokenType.SEMI,
103            TokenType.ANTI,
104            TokenType.SETTINGS,
105            TokenType.FORMAT,
106        }
107
108        LOG_DEFAULTS_TO_LN = True
109
110        QUERY_MODIFIER_PARSERS = {
111            **parser.Parser.QUERY_MODIFIER_PARSERS,
112            "settings": lambda self: self._parse_csv(self._parse_conjunction)
113            if self._match(TokenType.SETTINGS)
114            else None,
115            "format": lambda self: self._parse_id_var() if self._match(TokenType.FORMAT) else None,
116        }
117
118        def _parse_conjunction(self) -> t.Optional[exp.Expression]:
119            this = super()._parse_conjunction()
120
121            if self._match(TokenType.PLACEHOLDER):
122                return self.expression(
123                    exp.If,
124                    this=this,
125                    true=self._parse_conjunction(),
126                    false=self._match(TokenType.COLON) and self._parse_conjunction(),
127                )
128
129            return this
130
131        def _parse_placeholder(self) -> t.Optional[exp.Expression]:
132            """
133            Parse a placeholder expression like SELECT {abc: UInt32} or FROM {table: Identifier}
134            https://clickhouse.com/docs/en/sql-reference/syntax#defining-and-using-query-parameters
135            """
136            if not self._match(TokenType.L_BRACE):
137                return None
138
139            this = self._parse_id_var()
140            self._match(TokenType.COLON)
141            kind = self._parse_types(check_func=False) or (
142                self._match_text_seq("IDENTIFIER") and "Identifier"
143            )
144
145            if not kind:
146                self.raise_error("Expecting a placeholder type or 'Identifier' for tables")
147            elif not self._match(TokenType.R_BRACE):
148                self.raise_error("Expecting }")
149
150            return self.expression(exp.Placeholder, this=this, kind=kind)
151
152        def _parse_in(self, this: t.Optional[exp.Expression], is_global: bool = False) -> exp.In:
153            this = super()._parse_in(this)
154            this.set("is_global", is_global)
155            return this
156
157        def _parse_table(
158            self, schema: bool = False, alias_tokens: t.Optional[t.Collection[TokenType]] = None
159        ) -> t.Optional[exp.Expression]:
160            this = super()._parse_table(schema=schema, alias_tokens=alias_tokens)
161
162            if self._match(TokenType.FINAL):
163                this = self.expression(exp.Final, this=this)
164
165            return this
166
167        def _parse_position(self, haystack_first: bool = False) -> exp.StrPosition:
168            return super()._parse_position(haystack_first=True)
169
170        # https://clickhouse.com/docs/en/sql-reference/statements/select/with/
171        def _parse_cte(self) -> exp.CTE:
172            index = self._index
173            try:
174                # WITH <identifier> AS <subquery expression>
175                return super()._parse_cte()
176            except ParseError:
177                # WITH <expression> AS <identifier>
178                self._retreat(index)
179                statement = self._parse_statement()
180
181                if statement and isinstance(statement.this, exp.Alias):
182                    self.raise_error("Expected CTE to have alias")
183
184                return self.expression(exp.CTE, this=statement, alias=statement and statement.this)
185
186        def _parse_join_parts(
187            self,
188        ) -> t.Tuple[t.Optional[Token], t.Optional[Token], t.Optional[Token]]:
189            is_global = self._match(TokenType.GLOBAL) and self._prev
190            kind_pre = self._match_set(self.JOIN_KINDS, advance=False) and self._prev
191
192            if kind_pre:
193                kind = self._match_set(self.JOIN_KINDS) and self._prev
194                side = self._match_set(self.JOIN_SIDES) and self._prev
195                return is_global, side, kind
196
197            return (
198                is_global,
199                self._match_set(self.JOIN_SIDES) and self._prev,
200                self._match_set(self.JOIN_KINDS) and self._prev,
201            )
202
203        def _parse_join(self, skip_join_token: bool = False) -> t.Optional[exp.Join]:
204            join = super()._parse_join(skip_join_token)
205
206            if join:
207                join.set("global", join.args.pop("method", None))
208            return join
209
210        def _parse_function(
211            self, functions: t.Optional[t.Dict[str, t.Callable]] = None, anonymous: bool = False
212        ) -> t.Optional[exp.Expression]:
213            func = super()._parse_function(functions, anonymous)
214
215            if isinstance(func, exp.Anonymous):
216                params = self._parse_func_params(func)
217
218                if params:
219                    return self.expression(
220                        exp.ParameterizedAgg,
221                        this=func.this,
222                        expressions=func.expressions,
223                        params=params,
224                    )
225
226            return func
227
228        def _parse_func_params(
229            self, this: t.Optional[exp.Func] = None
230        ) -> t.Optional[t.List[t.Optional[exp.Expression]]]:
231            if self._match_pair(TokenType.R_PAREN, TokenType.L_PAREN):
232                return self._parse_csv(self._parse_lambda)
233
234            if self._match(TokenType.L_PAREN):
235                params = self._parse_csv(self._parse_lambda)
236                self._match_r_paren(this)
237                return params
238
239            return None
240
241        def _parse_quantile(self) -> exp.Quantile:
242            this = self._parse_lambda()
243            params = self._parse_func_params()
244            if params:
245                return self.expression(exp.Quantile, this=params[0], quantile=this)
246            return self.expression(exp.Quantile, this=this, quantile=exp.Literal.number(0.5))
247
248        def _parse_wrapped_id_vars(
249            self, optional: bool = False
250        ) -> t.List[t.Optional[exp.Expression]]:
251            return super()._parse_wrapped_id_vars(optional=True)
252
253        def _parse_primary_key(
254            self, wrapped_optional: bool = False, in_props: bool = False
255        ) -> exp.PrimaryKeyColumnConstraint | exp.PrimaryKey:
256            return super()._parse_primary_key(
257                wrapped_optional=wrapped_optional or in_props, in_props=in_props
258            )
259
260        def _parse_on_property(self) -> t.Optional[exp.Expression]:
261            index = self._index
262            if self._match_text_seq("CLUSTER"):
263                this = self._parse_id_var()
264                if this:
265                    return self.expression(exp.OnCluster, this=this)
266                else:
267                    self._retreat(index)
268            return None
269
270    class Generator(generator.Generator):
271        STRUCT_DELIMITER = ("(", ")")
272
273        TYPE_MAPPING = {
274            **generator.Generator.TYPE_MAPPING,
275            exp.DataType.Type.ARRAY: "Array",
276            exp.DataType.Type.BIGINT: "Int64",
277            exp.DataType.Type.DATETIME64: "DateTime64",
278            exp.DataType.Type.DOUBLE: "Float64",
279            exp.DataType.Type.FLOAT: "Float32",
280            exp.DataType.Type.INT: "Int32",
281            exp.DataType.Type.INT128: "Int128",
282            exp.DataType.Type.INT256: "Int256",
283            exp.DataType.Type.MAP: "Map",
284            exp.DataType.Type.NULLABLE: "Nullable",
285            exp.DataType.Type.SMALLINT: "Int16",
286            exp.DataType.Type.STRUCT: "Tuple",
287            exp.DataType.Type.TINYINT: "Int8",
288            exp.DataType.Type.UBIGINT: "UInt64",
289            exp.DataType.Type.UINT: "UInt32",
290            exp.DataType.Type.UINT128: "UInt128",
291            exp.DataType.Type.UINT256: "UInt256",
292            exp.DataType.Type.USMALLINT: "UInt16",
293            exp.DataType.Type.UTINYINT: "UInt8",
294        }
295
296        TRANSFORMS = {
297            **generator.Generator.TRANSFORMS,
298            exp.AnyValue: rename_func("any"),
299            exp.ApproxDistinct: rename_func("uniq"),
300            exp.Array: inline_array_sql,
301            exp.CastToStrType: rename_func("CAST"),
302            exp.Final: lambda self, e: f"{self.sql(e, 'this')} FINAL",
303            exp.Map: lambda self, e: _lower_func(var_map_sql(self, e)),
304            exp.PartitionedByProperty: lambda self, e: f"PARTITION BY {self.sql(e, 'this')}",
305            exp.Pivot: no_pivot_sql,
306            exp.Quantile: lambda self, e: self.func("quantile", e.args.get("quantile"))
307            + f"({self.sql(e, 'this')})",
308            exp.RegexpLike: lambda self, e: f"match({self.format_args(e.this, e.expression)})",
309            exp.StrPosition: lambda self, e: f"position({self.format_args(e.this, e.args.get('substr'), e.args.get('position'))})",
310            exp.VarMap: lambda self, e: _lower_func(var_map_sql(self, e)),
311        }
312
313        PROPERTIES_LOCATION = {
314            **generator.Generator.PROPERTIES_LOCATION,
315            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
316            exp.PartitionedByProperty: exp.Properties.Location.POST_SCHEMA,
317            exp.OnCluster: exp.Properties.Location.POST_NAME,
318        }
319
320        JOIN_HINTS = False
321        TABLE_HINTS = False
322        EXPLICIT_UNION = True
323        GROUPINGS_SEP = ""
324
325        # there's no list in docs, but it can be found in Clickhouse code
326        # see `ClickHouse/src/Parsers/ParserCreate*.cpp`
327        ON_CLUSTER_TARGETS = {
328            "DATABASE",
329            "TABLE",
330            "VIEW",
331            "DICTIONARY",
332            "INDEX",
333            "FUNCTION",
334            "NAMED COLLECTION",
335        }
336
337        def safeconcat_sql(self, expression: exp.SafeConcat) -> str:
338            # Clickhouse errors out if we try to cast a NULL value to TEXT
339            return self.func(
340                "CONCAT",
341                *[
342                    exp.func("if", e.is_(exp.null()), e, exp.cast(e, "text"))
343                    for e in expression.expressions
344                ],
345            )
346
347        def cte_sql(self, expression: exp.CTE) -> str:
348            if isinstance(expression.this, exp.Alias):
349                return self.sql(expression, "this")
350
351            return super().cte_sql(expression)
352
353        def after_limit_modifiers(self, expression: exp.Expression) -> t.List[str]:
354            return super().after_limit_modifiers(expression) + [
355                self.seg("SETTINGS ") + self.expressions(expression, key="settings", flat=True)
356                if expression.args.get("settings")
357                else "",
358                self.seg("FORMAT ") + self.sql(expression, "format")
359                if expression.args.get("format")
360                else "",
361            ]
362
363        def parameterizedagg_sql(self, expression: exp.Anonymous) -> str:
364            params = self.expressions(expression, "params", flat=True)
365            return self.func(expression.name, *expression.expressions) + f"({params})"
366
367        def placeholder_sql(self, expression: exp.Placeholder) -> str:
368            return f"{{{expression.name}: {self.sql(expression, 'kind')}}}"
369
370        def oncluster_sql(self, expression: exp.OnCluster) -> str:
371            return f"ON CLUSTER {self.sql(expression, 'this')}"
372
373        def createable_sql(
374            self,
375            expression: exp.Create,
376            locations: dict[exp.Properties.Location, list[exp.Property]],
377        ) -> str:
378            kind = self.sql(expression, "kind").upper()
379            if kind in self.ON_CLUSTER_TARGETS and locations.get(exp.Properties.Location.POST_NAME):
380                this_name = self.sql(expression.this, "this")
381                this_properties = " ".join(
382                    [self.sql(prop) for prop in locations[exp.Properties.Location.POST_NAME]]
383                )
384                this_schema = self.schema_columns_sql(expression.this)
385                return f"{this_name}{self.sep()}{this_properties}{self.sep()}{this_schema}"
386            return super().createable_sql(expression, locations)
class ClickHouse(sqlglot.dialects.dialect.Dialect):
 24class ClickHouse(Dialect):
 25    NORMALIZE_FUNCTIONS: bool | str = False
 26    NULL_ORDERING = "nulls_are_last"
 27    STRICT_STRING_CONCAT = True
 28
 29    class Tokenizer(tokens.Tokenizer):
 30        COMMENTS = ["--", "#", "#!", ("/*", "*/")]
 31        IDENTIFIERS = ['"', "`"]
 32        STRING_ESCAPES = ["'", "\\"]
 33        BIT_STRINGS = [("0b", "")]
 34        HEX_STRINGS = [("0x", ""), ("0X", "")]
 35
 36        KEYWORDS = {
 37            **tokens.Tokenizer.KEYWORDS,
 38            "ATTACH": TokenType.COMMAND,
 39            "DATETIME64": TokenType.DATETIME64,
 40            "DICTIONARY": TokenType.DICTIONARY,
 41            "FINAL": TokenType.FINAL,
 42            "FLOAT32": TokenType.FLOAT,
 43            "FLOAT64": TokenType.DOUBLE,
 44            "GLOBAL": TokenType.GLOBAL,
 45            "INT128": TokenType.INT128,
 46            "INT16": TokenType.SMALLINT,
 47            "INT256": TokenType.INT256,
 48            "INT32": TokenType.INT,
 49            "INT64": TokenType.BIGINT,
 50            "INT8": TokenType.TINYINT,
 51            "MAP": TokenType.MAP,
 52            "TUPLE": TokenType.STRUCT,
 53            "UINT128": TokenType.UINT128,
 54            "UINT16": TokenType.USMALLINT,
 55            "UINT256": TokenType.UINT256,
 56            "UINT32": TokenType.UINT,
 57            "UINT64": TokenType.UBIGINT,
 58            "UINT8": TokenType.UTINYINT,
 59        }
 60
 61    class Parser(parser.Parser):
 62        FUNCTIONS = {
 63            **parser.Parser.FUNCTIONS,
 64            "ANY": exp.AnyValue.from_arg_list,
 65            "MAP": parse_var_map,
 66            "MATCH": exp.RegexpLike.from_arg_list,
 67            "UNIQ": exp.ApproxDistinct.from_arg_list,
 68        }
 69
 70        FUNCTIONS_WITH_ALIASED_ARGS = {*parser.Parser.FUNCTIONS_WITH_ALIASED_ARGS, "TUPLE"}
 71
 72        FUNCTION_PARSERS = {
 73            **parser.Parser.FUNCTION_PARSERS,
 74            "QUANTILE": lambda self: self._parse_quantile(),
 75        }
 76
 77        FUNCTION_PARSERS.pop("MATCH")
 78
 79        NO_PAREN_FUNCTION_PARSERS = parser.Parser.NO_PAREN_FUNCTION_PARSERS.copy()
 80        NO_PAREN_FUNCTION_PARSERS.pop(TokenType.ANY)
 81
 82        RANGE_PARSERS = {
 83            **parser.Parser.RANGE_PARSERS,
 84            TokenType.GLOBAL: lambda self, this: self._match(TokenType.IN)
 85            and self._parse_in(this, is_global=True),
 86        }
 87
 88        # The PLACEHOLDER entry is popped because 1) it doesn't affect Clickhouse (it corresponds to
 89        # the postgres-specific JSONBContains parser) and 2) it makes parsing the ternary op simpler.
 90        COLUMN_OPERATORS = parser.Parser.COLUMN_OPERATORS.copy()
 91        COLUMN_OPERATORS.pop(TokenType.PLACEHOLDER)
 92
 93        JOIN_KINDS = {
 94            *parser.Parser.JOIN_KINDS,
 95            TokenType.ANY,
 96            TokenType.ASOF,
 97            TokenType.ANTI,
 98            TokenType.SEMI,
 99        }
100
101        TABLE_ALIAS_TOKENS = {*parser.Parser.TABLE_ALIAS_TOKENS} - {
102            TokenType.ANY,
103            TokenType.SEMI,
104            TokenType.ANTI,
105            TokenType.SETTINGS,
106            TokenType.FORMAT,
107        }
108
109        LOG_DEFAULTS_TO_LN = True
110
111        QUERY_MODIFIER_PARSERS = {
112            **parser.Parser.QUERY_MODIFIER_PARSERS,
113            "settings": lambda self: self._parse_csv(self._parse_conjunction)
114            if self._match(TokenType.SETTINGS)
115            else None,
116            "format": lambda self: self._parse_id_var() if self._match(TokenType.FORMAT) else None,
117        }
118
119        def _parse_conjunction(self) -> t.Optional[exp.Expression]:
120            this = super()._parse_conjunction()
121
122            if self._match(TokenType.PLACEHOLDER):
123                return self.expression(
124                    exp.If,
125                    this=this,
126                    true=self._parse_conjunction(),
127                    false=self._match(TokenType.COLON) and self._parse_conjunction(),
128                )
129
130            return this
131
132        def _parse_placeholder(self) -> t.Optional[exp.Expression]:
133            """
134            Parse a placeholder expression like SELECT {abc: UInt32} or FROM {table: Identifier}
135            https://clickhouse.com/docs/en/sql-reference/syntax#defining-and-using-query-parameters
136            """
137            if not self._match(TokenType.L_BRACE):
138                return None
139
140            this = self._parse_id_var()
141            self._match(TokenType.COLON)
142            kind = self._parse_types(check_func=False) or (
143                self._match_text_seq("IDENTIFIER") and "Identifier"
144            )
145
146            if not kind:
147                self.raise_error("Expecting a placeholder type or 'Identifier' for tables")
148            elif not self._match(TokenType.R_BRACE):
149                self.raise_error("Expecting }")
150
151            return self.expression(exp.Placeholder, this=this, kind=kind)
152
153        def _parse_in(self, this: t.Optional[exp.Expression], is_global: bool = False) -> exp.In:
154            this = super()._parse_in(this)
155            this.set("is_global", is_global)
156            return this
157
158        def _parse_table(
159            self, schema: bool = False, alias_tokens: t.Optional[t.Collection[TokenType]] = None
160        ) -> t.Optional[exp.Expression]:
161            this = super()._parse_table(schema=schema, alias_tokens=alias_tokens)
162
163            if self._match(TokenType.FINAL):
164                this = self.expression(exp.Final, this=this)
165
166            return this
167
168        def _parse_position(self, haystack_first: bool = False) -> exp.StrPosition:
169            return super()._parse_position(haystack_first=True)
170
171        # https://clickhouse.com/docs/en/sql-reference/statements/select/with/
172        def _parse_cte(self) -> exp.CTE:
173            index = self._index
174            try:
175                # WITH <identifier> AS <subquery expression>
176                return super()._parse_cte()
177            except ParseError:
178                # WITH <expression> AS <identifier>
179                self._retreat(index)
180                statement = self._parse_statement()
181
182                if statement and isinstance(statement.this, exp.Alias):
183                    self.raise_error("Expected CTE to have alias")
184
185                return self.expression(exp.CTE, this=statement, alias=statement and statement.this)
186
187        def _parse_join_parts(
188            self,
189        ) -> t.Tuple[t.Optional[Token], t.Optional[Token], t.Optional[Token]]:
190            is_global = self._match(TokenType.GLOBAL) and self._prev
191            kind_pre = self._match_set(self.JOIN_KINDS, advance=False) and self._prev
192
193            if kind_pre:
194                kind = self._match_set(self.JOIN_KINDS) and self._prev
195                side = self._match_set(self.JOIN_SIDES) and self._prev
196                return is_global, side, kind
197
198            return (
199                is_global,
200                self._match_set(self.JOIN_SIDES) and self._prev,
201                self._match_set(self.JOIN_KINDS) and self._prev,
202            )
203
204        def _parse_join(self, skip_join_token: bool = False) -> t.Optional[exp.Join]:
205            join = super()._parse_join(skip_join_token)
206
207            if join:
208                join.set("global", join.args.pop("method", None))
209            return join
210
211        def _parse_function(
212            self, functions: t.Optional[t.Dict[str, t.Callable]] = None, anonymous: bool = False
213        ) -> t.Optional[exp.Expression]:
214            func = super()._parse_function(functions, anonymous)
215
216            if isinstance(func, exp.Anonymous):
217                params = self._parse_func_params(func)
218
219                if params:
220                    return self.expression(
221                        exp.ParameterizedAgg,
222                        this=func.this,
223                        expressions=func.expressions,
224                        params=params,
225                    )
226
227            return func
228
229        def _parse_func_params(
230            self, this: t.Optional[exp.Func] = None
231        ) -> t.Optional[t.List[t.Optional[exp.Expression]]]:
232            if self._match_pair(TokenType.R_PAREN, TokenType.L_PAREN):
233                return self._parse_csv(self._parse_lambda)
234
235            if self._match(TokenType.L_PAREN):
236                params = self._parse_csv(self._parse_lambda)
237                self._match_r_paren(this)
238                return params
239
240            return None
241
242        def _parse_quantile(self) -> exp.Quantile:
243            this = self._parse_lambda()
244            params = self._parse_func_params()
245            if params:
246                return self.expression(exp.Quantile, this=params[0], quantile=this)
247            return self.expression(exp.Quantile, this=this, quantile=exp.Literal.number(0.5))
248
249        def _parse_wrapped_id_vars(
250            self, optional: bool = False
251        ) -> t.List[t.Optional[exp.Expression]]:
252            return super()._parse_wrapped_id_vars(optional=True)
253
254        def _parse_primary_key(
255            self, wrapped_optional: bool = False, in_props: bool = False
256        ) -> exp.PrimaryKeyColumnConstraint | exp.PrimaryKey:
257            return super()._parse_primary_key(
258                wrapped_optional=wrapped_optional or in_props, in_props=in_props
259            )
260
261        def _parse_on_property(self) -> t.Optional[exp.Expression]:
262            index = self._index
263            if self._match_text_seq("CLUSTER"):
264                this = self._parse_id_var()
265                if this:
266                    return self.expression(exp.OnCluster, this=this)
267                else:
268                    self._retreat(index)
269            return None
270
271    class Generator(generator.Generator):
272        STRUCT_DELIMITER = ("(", ")")
273
274        TYPE_MAPPING = {
275            **generator.Generator.TYPE_MAPPING,
276            exp.DataType.Type.ARRAY: "Array",
277            exp.DataType.Type.BIGINT: "Int64",
278            exp.DataType.Type.DATETIME64: "DateTime64",
279            exp.DataType.Type.DOUBLE: "Float64",
280            exp.DataType.Type.FLOAT: "Float32",
281            exp.DataType.Type.INT: "Int32",
282            exp.DataType.Type.INT128: "Int128",
283            exp.DataType.Type.INT256: "Int256",
284            exp.DataType.Type.MAP: "Map",
285            exp.DataType.Type.NULLABLE: "Nullable",
286            exp.DataType.Type.SMALLINT: "Int16",
287            exp.DataType.Type.STRUCT: "Tuple",
288            exp.DataType.Type.TINYINT: "Int8",
289            exp.DataType.Type.UBIGINT: "UInt64",
290            exp.DataType.Type.UINT: "UInt32",
291            exp.DataType.Type.UINT128: "UInt128",
292            exp.DataType.Type.UINT256: "UInt256",
293            exp.DataType.Type.USMALLINT: "UInt16",
294            exp.DataType.Type.UTINYINT: "UInt8",
295        }
296
297        TRANSFORMS = {
298            **generator.Generator.TRANSFORMS,
299            exp.AnyValue: rename_func("any"),
300            exp.ApproxDistinct: rename_func("uniq"),
301            exp.Array: inline_array_sql,
302            exp.CastToStrType: rename_func("CAST"),
303            exp.Final: lambda self, e: f"{self.sql(e, 'this')} FINAL",
304            exp.Map: lambda self, e: _lower_func(var_map_sql(self, e)),
305            exp.PartitionedByProperty: lambda self, e: f"PARTITION BY {self.sql(e, 'this')}",
306            exp.Pivot: no_pivot_sql,
307            exp.Quantile: lambda self, e: self.func("quantile", e.args.get("quantile"))
308            + f"({self.sql(e, 'this')})",
309            exp.RegexpLike: lambda self, e: f"match({self.format_args(e.this, e.expression)})",
310            exp.StrPosition: lambda self, e: f"position({self.format_args(e.this, e.args.get('substr'), e.args.get('position'))})",
311            exp.VarMap: lambda self, e: _lower_func(var_map_sql(self, e)),
312        }
313
314        PROPERTIES_LOCATION = {
315            **generator.Generator.PROPERTIES_LOCATION,
316            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
317            exp.PartitionedByProperty: exp.Properties.Location.POST_SCHEMA,
318            exp.OnCluster: exp.Properties.Location.POST_NAME,
319        }
320
321        JOIN_HINTS = False
322        TABLE_HINTS = False
323        EXPLICIT_UNION = True
324        GROUPINGS_SEP = ""
325
326        # there's no list in docs, but it can be found in Clickhouse code
327        # see `ClickHouse/src/Parsers/ParserCreate*.cpp`
328        ON_CLUSTER_TARGETS = {
329            "DATABASE",
330            "TABLE",
331            "VIEW",
332            "DICTIONARY",
333            "INDEX",
334            "FUNCTION",
335            "NAMED COLLECTION",
336        }
337
338        def safeconcat_sql(self, expression: exp.SafeConcat) -> str:
339            # Clickhouse errors out if we try to cast a NULL value to TEXT
340            return self.func(
341                "CONCAT",
342                *[
343                    exp.func("if", e.is_(exp.null()), e, exp.cast(e, "text"))
344                    for e in expression.expressions
345                ],
346            )
347
348        def cte_sql(self, expression: exp.CTE) -> str:
349            if isinstance(expression.this, exp.Alias):
350                return self.sql(expression, "this")
351
352            return super().cte_sql(expression)
353
354        def after_limit_modifiers(self, expression: exp.Expression) -> t.List[str]:
355            return super().after_limit_modifiers(expression) + [
356                self.seg("SETTINGS ") + self.expressions(expression, key="settings", flat=True)
357                if expression.args.get("settings")
358                else "",
359                self.seg("FORMAT ") + self.sql(expression, "format")
360                if expression.args.get("format")
361                else "",
362            ]
363
364        def parameterizedagg_sql(self, expression: exp.Anonymous) -> str:
365            params = self.expressions(expression, "params", flat=True)
366            return self.func(expression.name, *expression.expressions) + f"({params})"
367
368        def placeholder_sql(self, expression: exp.Placeholder) -> str:
369            return f"{{{expression.name}: {self.sql(expression, 'kind')}}}"
370
371        def oncluster_sql(self, expression: exp.OnCluster) -> str:
372            return f"ON CLUSTER {self.sql(expression, 'this')}"
373
374        def createable_sql(
375            self,
376            expression: exp.Create,
377            locations: dict[exp.Properties.Location, list[exp.Property]],
378        ) -> str:
379            kind = self.sql(expression, "kind").upper()
380            if kind in self.ON_CLUSTER_TARGETS and locations.get(exp.Properties.Location.POST_NAME):
381                this_name = self.sql(expression.this, "this")
382                this_properties = " ".join(
383                    [self.sql(prop) for prop in locations[exp.Properties.Location.POST_NAME]]
384                )
385                this_schema = self.schema_columns_sql(expression.this)
386                return f"{this_name}{self.sep()}{this_properties}{self.sep()}{this_schema}"
387            return super().createable_sql(expression, locations)
class ClickHouse.Tokenizer(sqlglot.tokens.Tokenizer):
29    class Tokenizer(tokens.Tokenizer):
30        COMMENTS = ["--", "#", "#!", ("/*", "*/")]
31        IDENTIFIERS = ['"', "`"]
32        STRING_ESCAPES = ["'", "\\"]
33        BIT_STRINGS = [("0b", "")]
34        HEX_STRINGS = [("0x", ""), ("0X", "")]
35
36        KEYWORDS = {
37            **tokens.Tokenizer.KEYWORDS,
38            "ATTACH": TokenType.COMMAND,
39            "DATETIME64": TokenType.DATETIME64,
40            "DICTIONARY": TokenType.DICTIONARY,
41            "FINAL": TokenType.FINAL,
42            "FLOAT32": TokenType.FLOAT,
43            "FLOAT64": TokenType.DOUBLE,
44            "GLOBAL": TokenType.GLOBAL,
45            "INT128": TokenType.INT128,
46            "INT16": TokenType.SMALLINT,
47            "INT256": TokenType.INT256,
48            "INT32": TokenType.INT,
49            "INT64": TokenType.BIGINT,
50            "INT8": TokenType.TINYINT,
51            "MAP": TokenType.MAP,
52            "TUPLE": TokenType.STRUCT,
53            "UINT128": TokenType.UINT128,
54            "UINT16": TokenType.USMALLINT,
55            "UINT256": TokenType.UINT256,
56            "UINT32": TokenType.UINT,
57            "UINT64": TokenType.UBIGINT,
58            "UINT8": TokenType.UTINYINT,
59        }
class ClickHouse.Parser(sqlglot.parser.Parser):
 61    class Parser(parser.Parser):
 62        FUNCTIONS = {
 63            **parser.Parser.FUNCTIONS,
 64            "ANY": exp.AnyValue.from_arg_list,
 65            "MAP": parse_var_map,
 66            "MATCH": exp.RegexpLike.from_arg_list,
 67            "UNIQ": exp.ApproxDistinct.from_arg_list,
 68        }
 69
 70        FUNCTIONS_WITH_ALIASED_ARGS = {*parser.Parser.FUNCTIONS_WITH_ALIASED_ARGS, "TUPLE"}
 71
 72        FUNCTION_PARSERS = {
 73            **parser.Parser.FUNCTION_PARSERS,
 74            "QUANTILE": lambda self: self._parse_quantile(),
 75        }
 76
 77        FUNCTION_PARSERS.pop("MATCH")
 78
 79        NO_PAREN_FUNCTION_PARSERS = parser.Parser.NO_PAREN_FUNCTION_PARSERS.copy()
 80        NO_PAREN_FUNCTION_PARSERS.pop(TokenType.ANY)
 81
 82        RANGE_PARSERS = {
 83            **parser.Parser.RANGE_PARSERS,
 84            TokenType.GLOBAL: lambda self, this: self._match(TokenType.IN)
 85            and self._parse_in(this, is_global=True),
 86        }
 87
 88        # The PLACEHOLDER entry is popped because 1) it doesn't affect Clickhouse (it corresponds to
 89        # the postgres-specific JSONBContains parser) and 2) it makes parsing the ternary op simpler.
 90        COLUMN_OPERATORS = parser.Parser.COLUMN_OPERATORS.copy()
 91        COLUMN_OPERATORS.pop(TokenType.PLACEHOLDER)
 92
 93        JOIN_KINDS = {
 94            *parser.Parser.JOIN_KINDS,
 95            TokenType.ANY,
 96            TokenType.ASOF,
 97            TokenType.ANTI,
 98            TokenType.SEMI,
 99        }
100
101        TABLE_ALIAS_TOKENS = {*parser.Parser.TABLE_ALIAS_TOKENS} - {
102            TokenType.ANY,
103            TokenType.SEMI,
104            TokenType.ANTI,
105            TokenType.SETTINGS,
106            TokenType.FORMAT,
107        }
108
109        LOG_DEFAULTS_TO_LN = True
110
111        QUERY_MODIFIER_PARSERS = {
112            **parser.Parser.QUERY_MODIFIER_PARSERS,
113            "settings": lambda self: self._parse_csv(self._parse_conjunction)
114            if self._match(TokenType.SETTINGS)
115            else None,
116            "format": lambda self: self._parse_id_var() if self._match(TokenType.FORMAT) else None,
117        }
118
119        def _parse_conjunction(self) -> t.Optional[exp.Expression]:
120            this = super()._parse_conjunction()
121
122            if self._match(TokenType.PLACEHOLDER):
123                return self.expression(
124                    exp.If,
125                    this=this,
126                    true=self._parse_conjunction(),
127                    false=self._match(TokenType.COLON) and self._parse_conjunction(),
128                )
129
130            return this
131
132        def _parse_placeholder(self) -> t.Optional[exp.Expression]:
133            """
134            Parse a placeholder expression like SELECT {abc: UInt32} or FROM {table: Identifier}
135            https://clickhouse.com/docs/en/sql-reference/syntax#defining-and-using-query-parameters
136            """
137            if not self._match(TokenType.L_BRACE):
138                return None
139
140            this = self._parse_id_var()
141            self._match(TokenType.COLON)
142            kind = self._parse_types(check_func=False) or (
143                self._match_text_seq("IDENTIFIER") and "Identifier"
144            )
145
146            if not kind:
147                self.raise_error("Expecting a placeholder type or 'Identifier' for tables")
148            elif not self._match(TokenType.R_BRACE):
149                self.raise_error("Expecting }")
150
151            return self.expression(exp.Placeholder, this=this, kind=kind)
152
153        def _parse_in(self, this: t.Optional[exp.Expression], is_global: bool = False) -> exp.In:
154            this = super()._parse_in(this)
155            this.set("is_global", is_global)
156            return this
157
158        def _parse_table(
159            self, schema: bool = False, alias_tokens: t.Optional[t.Collection[TokenType]] = None
160        ) -> t.Optional[exp.Expression]:
161            this = super()._parse_table(schema=schema, alias_tokens=alias_tokens)
162
163            if self._match(TokenType.FINAL):
164                this = self.expression(exp.Final, this=this)
165
166            return this
167
168        def _parse_position(self, haystack_first: bool = False) -> exp.StrPosition:
169            return super()._parse_position(haystack_first=True)
170
171        # https://clickhouse.com/docs/en/sql-reference/statements/select/with/
172        def _parse_cte(self) -> exp.CTE:
173            index = self._index
174            try:
175                # WITH <identifier> AS <subquery expression>
176                return super()._parse_cte()
177            except ParseError:
178                # WITH <expression> AS <identifier>
179                self._retreat(index)
180                statement = self._parse_statement()
181
182                if statement and isinstance(statement.this, exp.Alias):
183                    self.raise_error("Expected CTE to have alias")
184
185                return self.expression(exp.CTE, this=statement, alias=statement and statement.this)
186
187        def _parse_join_parts(
188            self,
189        ) -> t.Tuple[t.Optional[Token], t.Optional[Token], t.Optional[Token]]:
190            is_global = self._match(TokenType.GLOBAL) and self._prev
191            kind_pre = self._match_set(self.JOIN_KINDS, advance=False) and self._prev
192
193            if kind_pre:
194                kind = self._match_set(self.JOIN_KINDS) and self._prev
195                side = self._match_set(self.JOIN_SIDES) and self._prev
196                return is_global, side, kind
197
198            return (
199                is_global,
200                self._match_set(self.JOIN_SIDES) and self._prev,
201                self._match_set(self.JOIN_KINDS) and self._prev,
202            )
203
204        def _parse_join(self, skip_join_token: bool = False) -> t.Optional[exp.Join]:
205            join = super()._parse_join(skip_join_token)
206
207            if join:
208                join.set("global", join.args.pop("method", None))
209            return join
210
211        def _parse_function(
212            self, functions: t.Optional[t.Dict[str, t.Callable]] = None, anonymous: bool = False
213        ) -> t.Optional[exp.Expression]:
214            func = super()._parse_function(functions, anonymous)
215
216            if isinstance(func, exp.Anonymous):
217                params = self._parse_func_params(func)
218
219                if params:
220                    return self.expression(
221                        exp.ParameterizedAgg,
222                        this=func.this,
223                        expressions=func.expressions,
224                        params=params,
225                    )
226
227            return func
228
229        def _parse_func_params(
230            self, this: t.Optional[exp.Func] = None
231        ) -> t.Optional[t.List[t.Optional[exp.Expression]]]:
232            if self._match_pair(TokenType.R_PAREN, TokenType.L_PAREN):
233                return self._parse_csv(self._parse_lambda)
234
235            if self._match(TokenType.L_PAREN):
236                params = self._parse_csv(self._parse_lambda)
237                self._match_r_paren(this)
238                return params
239
240            return None
241
242        def _parse_quantile(self) -> exp.Quantile:
243            this = self._parse_lambda()
244            params = self._parse_func_params()
245            if params:
246                return self.expression(exp.Quantile, this=params[0], quantile=this)
247            return self.expression(exp.Quantile, this=this, quantile=exp.Literal.number(0.5))
248
249        def _parse_wrapped_id_vars(
250            self, optional: bool = False
251        ) -> t.List[t.Optional[exp.Expression]]:
252            return super()._parse_wrapped_id_vars(optional=True)
253
254        def _parse_primary_key(
255            self, wrapped_optional: bool = False, in_props: bool = False
256        ) -> exp.PrimaryKeyColumnConstraint | exp.PrimaryKey:
257            return super()._parse_primary_key(
258                wrapped_optional=wrapped_optional or in_props, in_props=in_props
259            )
260
261        def _parse_on_property(self) -> t.Optional[exp.Expression]:
262            index = self._index
263            if self._match_text_seq("CLUSTER"):
264                this = self._parse_id_var()
265                if this:
266                    return self.expression(exp.OnCluster, this=this)
267                else:
268                    self._retreat(index)
269            return None

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: Determines 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
class ClickHouse.Generator(sqlglot.generator.Generator):
271    class Generator(generator.Generator):
272        STRUCT_DELIMITER = ("(", ")")
273
274        TYPE_MAPPING = {
275            **generator.Generator.TYPE_MAPPING,
276            exp.DataType.Type.ARRAY: "Array",
277            exp.DataType.Type.BIGINT: "Int64",
278            exp.DataType.Type.DATETIME64: "DateTime64",
279            exp.DataType.Type.DOUBLE: "Float64",
280            exp.DataType.Type.FLOAT: "Float32",
281            exp.DataType.Type.INT: "Int32",
282            exp.DataType.Type.INT128: "Int128",
283            exp.DataType.Type.INT256: "Int256",
284            exp.DataType.Type.MAP: "Map",
285            exp.DataType.Type.NULLABLE: "Nullable",
286            exp.DataType.Type.SMALLINT: "Int16",
287            exp.DataType.Type.STRUCT: "Tuple",
288            exp.DataType.Type.TINYINT: "Int8",
289            exp.DataType.Type.UBIGINT: "UInt64",
290            exp.DataType.Type.UINT: "UInt32",
291            exp.DataType.Type.UINT128: "UInt128",
292            exp.DataType.Type.UINT256: "UInt256",
293            exp.DataType.Type.USMALLINT: "UInt16",
294            exp.DataType.Type.UTINYINT: "UInt8",
295        }
296
297        TRANSFORMS = {
298            **generator.Generator.TRANSFORMS,
299            exp.AnyValue: rename_func("any"),
300            exp.ApproxDistinct: rename_func("uniq"),
301            exp.Array: inline_array_sql,
302            exp.CastToStrType: rename_func("CAST"),
303            exp.Final: lambda self, e: f"{self.sql(e, 'this')} FINAL",
304            exp.Map: lambda self, e: _lower_func(var_map_sql(self, e)),
305            exp.PartitionedByProperty: lambda self, e: f"PARTITION BY {self.sql(e, 'this')}",
306            exp.Pivot: no_pivot_sql,
307            exp.Quantile: lambda self, e: self.func("quantile", e.args.get("quantile"))
308            + f"({self.sql(e, 'this')})",
309            exp.RegexpLike: lambda self, e: f"match({self.format_args(e.this, e.expression)})",
310            exp.StrPosition: lambda self, e: f"position({self.format_args(e.this, e.args.get('substr'), e.args.get('position'))})",
311            exp.VarMap: lambda self, e: _lower_func(var_map_sql(self, e)),
312        }
313
314        PROPERTIES_LOCATION = {
315            **generator.Generator.PROPERTIES_LOCATION,
316            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
317            exp.PartitionedByProperty: exp.Properties.Location.POST_SCHEMA,
318            exp.OnCluster: exp.Properties.Location.POST_NAME,
319        }
320
321        JOIN_HINTS = False
322        TABLE_HINTS = False
323        EXPLICIT_UNION = True
324        GROUPINGS_SEP = ""
325
326        # there's no list in docs, but it can be found in Clickhouse code
327        # see `ClickHouse/src/Parsers/ParserCreate*.cpp`
328        ON_CLUSTER_TARGETS = {
329            "DATABASE",
330            "TABLE",
331            "VIEW",
332            "DICTIONARY",
333            "INDEX",
334            "FUNCTION",
335            "NAMED COLLECTION",
336        }
337
338        def safeconcat_sql(self, expression: exp.SafeConcat) -> str:
339            # Clickhouse errors out if we try to cast a NULL value to TEXT
340            return self.func(
341                "CONCAT",
342                *[
343                    exp.func("if", e.is_(exp.null()), e, exp.cast(e, "text"))
344                    for e in expression.expressions
345                ],
346            )
347
348        def cte_sql(self, expression: exp.CTE) -> str:
349            if isinstance(expression.this, exp.Alias):
350                return self.sql(expression, "this")
351
352            return super().cte_sql(expression)
353
354        def after_limit_modifiers(self, expression: exp.Expression) -> t.List[str]:
355            return super().after_limit_modifiers(expression) + [
356                self.seg("SETTINGS ") + self.expressions(expression, key="settings", flat=True)
357                if expression.args.get("settings")
358                else "",
359                self.seg("FORMAT ") + self.sql(expression, "format")
360                if expression.args.get("format")
361                else "",
362            ]
363
364        def parameterizedagg_sql(self, expression: exp.Anonymous) -> str:
365            params = self.expressions(expression, "params", flat=True)
366            return self.func(expression.name, *expression.expressions) + f"({params})"
367
368        def placeholder_sql(self, expression: exp.Placeholder) -> str:
369            return f"{{{expression.name}: {self.sql(expression, 'kind')}}}"
370
371        def oncluster_sql(self, expression: exp.OnCluster) -> str:
372            return f"ON CLUSTER {self.sql(expression, 'this')}"
373
374        def createable_sql(
375            self,
376            expression: exp.Create,
377            locations: dict[exp.Properties.Location, list[exp.Property]],
378        ) -> str:
379            kind = self.sql(expression, "kind").upper()
380            if kind in self.ON_CLUSTER_TARGETS and locations.get(exp.Properties.Location.POST_NAME):
381                this_name = self.sql(expression.this, "this")
382                this_properties = " ".join(
383                    [self.sql(prop) for prop in locations[exp.Properties.Location.POST_NAME]]
384                )
385                this_schema = self.schema_columns_sql(expression.this)
386                return f"{this_name}{self.sep()}{this_properties}{self.sep()}{this_schema}"
387            return super().createable_sql(expression, locations)

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
def safeconcat_sql(self, expression: sqlglot.expressions.SafeConcat) -> str:
338        def safeconcat_sql(self, expression: exp.SafeConcat) -> str:
339            # Clickhouse errors out if we try to cast a NULL value to TEXT
340            return self.func(
341                "CONCAT",
342                *[
343                    exp.func("if", e.is_(exp.null()), e, exp.cast(e, "text"))
344                    for e in expression.expressions
345                ],
346            )
def cte_sql(self, expression: sqlglot.expressions.CTE) -> str:
348        def cte_sql(self, expression: exp.CTE) -> str:
349            if isinstance(expression.this, exp.Alias):
350                return self.sql(expression, "this")
351
352            return super().cte_sql(expression)
def after_limit_modifiers(self, expression: sqlglot.expressions.Expression) -> List[str]:
354        def after_limit_modifiers(self, expression: exp.Expression) -> t.List[str]:
355            return super().after_limit_modifiers(expression) + [
356                self.seg("SETTINGS ") + self.expressions(expression, key="settings", flat=True)
357                if expression.args.get("settings")
358                else "",
359                self.seg("FORMAT ") + self.sql(expression, "format")
360                if expression.args.get("format")
361                else "",
362            ]
def parameterizedagg_sql(self, expression: sqlglot.expressions.Anonymous) -> str:
364        def parameterizedagg_sql(self, expression: exp.Anonymous) -> str:
365            params = self.expressions(expression, "params", flat=True)
366            return self.func(expression.name, *expression.expressions) + f"({params})"
def placeholder_sql(self, expression: sqlglot.expressions.Placeholder) -> str:
368        def placeholder_sql(self, expression: exp.Placeholder) -> str:
369            return f"{{{expression.name}: {self.sql(expression, 'kind')}}}"
def oncluster_sql(self, expression: sqlglot.expressions.OnCluster) -> str:
371        def oncluster_sql(self, expression: exp.OnCluster) -> str:
372            return f"ON CLUSTER {self.sql(expression, 'this')}"
def createable_sql( self, expression: sqlglot.expressions.Create, locations: dict[sqlglot.expressions.Properties.Location, list[sqlglot.expressions.Property]]) -> str:
374        def createable_sql(
375            self,
376            expression: exp.Create,
377            locations: dict[exp.Properties.Location, list[exp.Property]],
378        ) -> str:
379            kind = self.sql(expression, "kind").upper()
380            if kind in self.ON_CLUSTER_TARGETS and locations.get(exp.Properties.Location.POST_NAME):
381                this_name = self.sql(expression.this, "this")
382                this_properties = " ".join(
383                    [self.sql(prop) for prop in locations[exp.Properties.Location.POST_NAME]]
384                )
385                this_schema = self.schema_columns_sql(expression.this)
386                return f"{this_name}{self.sep()}{this_properties}{self.sep()}{this_schema}"
387            return super().createable_sql(expression, locations)
Inherited Members
sqlglot.generator.Generator
Generator
generate
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
autoincrementcolumnconstraint_sql
compresscolumnconstraint_sql
generatedasidentitycolumnconstraint_sql
notnullcolumnconstraint_sql
primarykeycolumnconstraint_sql
uniquecolumnconstraint_sql
create_sql
clone_sql
describe_sql
prepend_ctes
with_sql
tablealias_sql
bitstring_sql
hexstring_sql
bytestring_sql
rawstring_sql
datatypesize_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_sql
likeproperty_sql
fallbackproperty_sql
journalproperty_sql
freespaceproperty_sql
checksumproperty_sql
mergeblockratioproperty_sql
datablocksizeproperty_sql
blockcompressionproperty_sql
isolatedloadingproperty_sql
lockingproperty_sql
withdataproperty_sql
insert_sql
intersect_sql
intersect_op
introducer_sql
pseudotype_sql
onconflict_sql
returning_sql
rowformatdelimitedproperty_sql
table_sql
tablesample_sql
pivot_sql
tuple_sql
update_sql
values_sql
var_sql
into_sql
from_sql
group_sql
having_sql
join_sql
lambda_sql
lateral_sql
limit_sql
offset_sql
setitem_sql
set_sql
pragma_sql
lock_sql
literal_sql
loaddata_sql
null_sql
boolean_sql
order_sql
cluster_sql
distribute_sql
sort_sql
ordered_sql
matchrecognize_sql
query_modifiers
offset_limit_modifiers
after_having_modifiers
select_sql
schema_sql
schema_columns_sql
star_sql
parameter_sql
sessionparameter_sql
subquery_sql
qualify_sql
union_sql
union_op
unnest_sql
where_sql
window_sql
partition_by_sql
windowspec_sql
withingroup_sql
between_sql
bracket_sql
all_sql
any_sql
exists_sql
case_sql
constraint_sql
nextvaluefor_sql
extract_sql
trim_sql
check_sql
foreignkey_sql
primarykey_sql
if_sql
matchagainst_sql
jsonkeyvalue_sql
jsonobject_sql
openjsoncolumndef_sql
openjson_sql
in_sql
in_unnest_op
interval_sql
return_sql
reference_sql
anonymous_sql
paren_sql
neg_sql
not_sql
alias_sql
aliases_sql
attimezone_sql
add_sql
and_sql
connector_sql
bitwiseand_sql
bitwiseleftshift_sql
bitwisenot_sql
bitwiseor_sql
bitwiserightshift_sql
bitwisexor_sql
cast_sql
currentdate_sql
collate_sql
command_sql
comment_sql
mergetreettlaction_sql
mergetreettl_sql
transaction_sql
commit_sql
rollback_sql
altercolumn_sql
renametable_sql
altertable_sql
droppartition_sql
addconstraint_sql
distinct_sql
ignorenulls_sql
respectnulls_sql
intdiv_sql
dpipe_sql
safedpipe_sql
div_sql
overlaps_sql
distance_sql
dot_sql
eq_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
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