Supports BigQuery Standard SQL.
1"""Supports BigQuery Standard SQL.""" 2 3from __future__ import annotations 4 5import typing as t 6 7from sqlglot import exp, generator, parser, tokens, transforms 8from sqlglot.dialects.dialect import ( 9 Dialect, 10 datestrtodate_sql, 11 inline_array_sql, 12 no_ilike_sql, 13 rename_func, 14 timestrtotime_sql, 15) 16from sqlglot.helper import seq_get 17from sqlglot.tokens import TokenType 18 19E = t.TypeVar("E", bound=exp.Expression) 20 21 22def _date_add(expression_class: t.Type[E]) -> t.Callable[[t.Sequence], E]: 23 def func(args): 24 interval = seq_get(args, 1) 25 return expression_class( 26 this=seq_get(args, 0), 27 expression=interval.this, 28 unit=interval.args.get("unit"), 29 ) 30 31 return func 32 33 34def _date_add_sql( 35 data_type: str, kind: str 36) -> t.Callable[[generator.Generator, exp.Expression], str]: 37 def func(self, expression): 38 this = self.sql(expression, "this") 39 return f"{data_type}_{kind}({this}, {self.sql(exp.Interval(this=expression.expression, unit=expression.args.get('unit') or exp.Literal.string('day')))})" 40 41 return func 42 43 44def _derived_table_values_to_unnest(self: generator.Generator, expression: exp.Values) -> str: 45 if not isinstance(expression.unnest().parent, exp.From): 46 expression = t.cast(exp.Values, transforms.remove_precision_parameterized_types(expression)) 47 return self.values_sql(expression) 48 rows = [tuple_exp.expressions for tuple_exp in expression.find_all(exp.Tuple)] 49 structs = [] 50 for row in rows: 51 aliases = [ 52 exp.alias_(value, column_name) 53 for value, column_name in zip(row, expression.args["alias"].args["columns"]) 54 ] 55 structs.append(exp.Struct(expressions=aliases)) 56 unnest_exp = exp.Unnest(expressions=[exp.Array(expressions=structs)]) 57 return self.unnest_sql(unnest_exp) 58 59 60def _returnsproperty_sql(self: generator.Generator, expression: exp.ReturnsProperty) -> str: 61 this = expression.this 62 if isinstance(this, exp.Schema): 63 this = f"{this.this} <{self.expressions(this)}>" 64 else: 65 this = self.sql(this) 66 return f"RETURNS {this}" 67 68 69def _create_sql(self: generator.Generator, expression: exp.Create) -> str: 70 kind = expression.args["kind"] 71 returns = expression.find(exp.ReturnsProperty) 72 if kind.upper() == "FUNCTION" and returns and returns.args.get("is_table"): 73 expression = expression.copy() 74 expression.set("kind", "TABLE FUNCTION") 75 if isinstance( 76 expression.expression, 77 ( 78 exp.Subquery, 79 exp.Literal, 80 ), 81 ): 82 expression.set("expression", expression.expression.this) 83 84 return self.create_sql(expression) 85 86 return self.create_sql(expression) 87 88 89def _unqualify_unnest(expression: exp.Expression) -> exp.Expression: 90 """Remove references to unnest table aliases since bigquery doesn't allow them. 91 92 These are added by the optimizer's qualify_column step. 93 """ 94 if isinstance(expression, exp.Select): 95 unnests = { 96 unnest.alias 97 for unnest in expression.args.get("from", exp.From(expressions=[])).expressions 98 if isinstance(unnest, exp.Unnest) and unnest.alias 99 } 100 101 if unnests: 102 expression = expression.copy() 103 104 for select in expression.expressions: 105 for column in select.find_all(exp.Column): 106 if column.table in unnests: 107 column.set("table", None) 108 109 return expression 110 111 112class BigQuery(Dialect): 113 unnest_column_only = True 114 time_mapping = { 115 "%M": "%-M", 116 "%d": "%-d", 117 "%m": "%-m", 118 "%y": "%-y", 119 "%H": "%-H", 120 "%I": "%-I", 121 "%S": "%-S", 122 "%j": "%-j", 123 } 124 125 class Tokenizer(tokens.Tokenizer): 126 QUOTES = [ 127 (prefix + quote, quote) if prefix else quote 128 for quote in ["'", '"', '"""', "'''"] 129 for prefix in ["", "r", "R"] 130 ] 131 COMMENTS = ["--", "#", ("/*", "*/")] 132 IDENTIFIERS = ["`"] 133 STRING_ESCAPES = ["\\"] 134 HEX_STRINGS = [("0x", ""), ("0X", "")] 135 136 KEYWORDS = { 137 **tokens.Tokenizer.KEYWORDS, 138 "BEGIN": TokenType.COMMAND, 139 "BEGIN TRANSACTION": TokenType.BEGIN, 140 "CURRENT_DATETIME": TokenType.CURRENT_DATETIME, 141 "CURRENT_TIME": TokenType.CURRENT_TIME, 142 "DECLARE": TokenType.COMMAND, 143 "GEOGRAPHY": TokenType.GEOGRAPHY, 144 "FLOAT64": TokenType.DOUBLE, 145 "INT64": TokenType.BIGINT, 146 "NOT DETERMINISTIC": TokenType.VOLATILE, 147 "UNKNOWN": TokenType.NULL, 148 } 149 KEYWORDS.pop("DIV") 150 151 class Parser(parser.Parser): 152 FUNCTIONS = { 153 **parser.Parser.FUNCTIONS, # type: ignore 154 "DATE_TRUNC": lambda args: exp.DateTrunc( 155 unit=exp.Literal.string(seq_get(args, 1).name), # type: ignore 156 this=seq_get(args, 0), 157 ), 158 "DATE_ADD": _date_add(exp.DateAdd), 159 "DATETIME_ADD": _date_add(exp.DatetimeAdd), 160 "DIV": lambda args: exp.IntDiv(this=seq_get(args, 0), expression=seq_get(args, 1)), 161 "REGEXP_CONTAINS": exp.RegexpLike.from_arg_list, 162 "TIME_ADD": _date_add(exp.TimeAdd), 163 "TIMESTAMP_ADD": _date_add(exp.TimestampAdd), 164 "DATE_SUB": _date_add(exp.DateSub), 165 "DATETIME_SUB": _date_add(exp.DatetimeSub), 166 "TIME_SUB": _date_add(exp.TimeSub), 167 "TIMESTAMP_SUB": _date_add(exp.TimestampSub), 168 "PARSE_TIMESTAMP": lambda args: exp.StrToTime( 169 this=seq_get(args, 1), format=seq_get(args, 0) 170 ), 171 } 172 173 FUNCTION_PARSERS = { 174 **parser.Parser.FUNCTION_PARSERS, # type: ignore 175 "ARRAY": lambda self: self.expression(exp.Array, expressions=[self._parse_statement()]), 176 } 177 FUNCTION_PARSERS.pop("TRIM") 178 179 NO_PAREN_FUNCTIONS = { 180 **parser.Parser.NO_PAREN_FUNCTIONS, # type: ignore 181 TokenType.CURRENT_DATETIME: exp.CurrentDatetime, 182 TokenType.CURRENT_TIME: exp.CurrentTime, 183 } 184 185 NESTED_TYPE_TOKENS = { 186 *parser.Parser.NESTED_TYPE_TOKENS, # type: ignore 187 TokenType.TABLE, 188 } 189 190 ID_VAR_TOKENS = { 191 *parser.Parser.ID_VAR_TOKENS, # type: ignore 192 TokenType.VALUES, 193 } 194 195 PROPERTY_PARSERS = { 196 **parser.Parser.PROPERTY_PARSERS, # type: ignore 197 "NOT DETERMINISTIC": lambda self: self.expression( 198 exp.VolatilityProperty, this=exp.Literal.string("VOLATILE") 199 ), 200 } 201 202 class Generator(generator.Generator): 203 TRANSFORMS = { 204 **generator.Generator.TRANSFORMS, # type: ignore 205 **transforms.REMOVE_PRECISION_PARAMETERIZED_TYPES, # type: ignore 206 exp.ArraySize: rename_func("ARRAY_LENGTH"), 207 exp.DateAdd: _date_add_sql("DATE", "ADD"), 208 exp.DateSub: _date_add_sql("DATE", "SUB"), 209 exp.DatetimeAdd: _date_add_sql("DATETIME", "ADD"), 210 exp.DatetimeSub: _date_add_sql("DATETIME", "SUB"), 211 exp.DateDiff: lambda self, e: f"DATE_DIFF({self.sql(e, 'this')}, {self.sql(e, 'expression')}, {self.sql(e.args.get('unit', 'DAY'))})", 212 exp.DateStrToDate: datestrtodate_sql, 213 exp.DateTrunc: lambda self, e: self.func("DATE_TRUNC", e.this, e.text("unit")), 214 exp.GroupConcat: rename_func("STRING_AGG"), 215 exp.ILike: no_ilike_sql, 216 exp.IntDiv: rename_func("DIV"), 217 exp.Select: transforms.preprocess( 218 [_unqualify_unnest], transforms.delegate("select_sql") 219 ), 220 exp.StrToTime: lambda self, e: f"PARSE_TIMESTAMP({self.format_time(e)}, {self.sql(e, 'this')})", 221 exp.TimeAdd: _date_add_sql("TIME", "ADD"), 222 exp.TimeSub: _date_add_sql("TIME", "SUB"), 223 exp.TimestampAdd: _date_add_sql("TIMESTAMP", "ADD"), 224 exp.TimestampSub: _date_add_sql("TIMESTAMP", "SUB"), 225 exp.TimeStrToTime: timestrtotime_sql, 226 exp.PartitionedByProperty: lambda self, e: f"PARTITION BY {self.sql(e, 'this')}", 227 exp.VariancePop: rename_func("VAR_POP"), 228 exp.Values: _derived_table_values_to_unnest, 229 exp.ReturnsProperty: _returnsproperty_sql, 230 exp.Create: _create_sql, 231 exp.Trim: lambda self, e: self.func(f"TRIM", e.this, e.expression), 232 exp.VolatilityProperty: lambda self, e: f"DETERMINISTIC" 233 if e.name == "IMMUTABLE" 234 else "NOT DETERMINISTIC", 235 exp.RegexpLike: rename_func("REGEXP_CONTAINS"), 236 } 237 238 TYPE_MAPPING = { 239 **generator.Generator.TYPE_MAPPING, # type: ignore 240 exp.DataType.Type.TINYINT: "INT64", 241 exp.DataType.Type.SMALLINT: "INT64", 242 exp.DataType.Type.INT: "INT64", 243 exp.DataType.Type.BIGINT: "INT64", 244 exp.DataType.Type.DECIMAL: "NUMERIC", 245 exp.DataType.Type.FLOAT: "FLOAT64", 246 exp.DataType.Type.DOUBLE: "FLOAT64", 247 exp.DataType.Type.BOOLEAN: "BOOL", 248 exp.DataType.Type.TEXT: "STRING", 249 exp.DataType.Type.VARCHAR: "STRING", 250 exp.DataType.Type.NVARCHAR: "STRING", 251 } 252 PROPERTIES_LOCATION = { 253 **generator.Generator.PROPERTIES_LOCATION, # type: ignore 254 exp.PartitionedByProperty: exp.Properties.Location.POST_SCHEMA, 255 } 256 257 EXPLICIT_UNION = True 258 259 def array_sql(self, expression: exp.Array) -> str: 260 first_arg = seq_get(expression.expressions, 0) 261 if isinstance(first_arg, exp.Subqueryable): 262 return f"ARRAY{self.wrap(self.sql(first_arg))}" 263 264 return inline_array_sql(self, expression) 265 266 def transaction_sql(self, *_) -> str: 267 return "BEGIN TRANSACTION" 268 269 def commit_sql(self, *_) -> str: 270 return "COMMIT TRANSACTION" 271 272 def rollback_sql(self, *_) -> str: 273 return "ROLLBACK TRANSACTION" 274 275 def in_unnest_op(self, expression: exp.Unnest) -> str: 276 return self.sql(expression) 277 278 def except_op(self, expression: exp.Except) -> str: 279 if not expression.args.get("distinct", False): 280 self.unsupported("EXCEPT without DISTINCT is not supported in BigQuery") 281 return f"EXCEPT{' DISTINCT' if expression.args.get('distinct') else ' ALL'}" 282 283 def intersect_op(self, expression: exp.Intersect) -> str: 284 if not expression.args.get("distinct", False): 285 self.unsupported("INTERSECT without DISTINCT is not supported in BigQuery") 286 return f"INTERSECT{' DISTINCT' if expression.args.get('distinct') else ' ALL'}"
113class BigQuery(Dialect): 114 unnest_column_only = True 115 time_mapping = { 116 "%M": "%-M", 117 "%d": "%-d", 118 "%m": "%-m", 119 "%y": "%-y", 120 "%H": "%-H", 121 "%I": "%-I", 122 "%S": "%-S", 123 "%j": "%-j", 124 } 125 126 class Tokenizer(tokens.Tokenizer): 127 QUOTES = [ 128 (prefix + quote, quote) if prefix else quote 129 for quote in ["'", '"', '"""', "'''"] 130 for prefix in ["", "r", "R"] 131 ] 132 COMMENTS = ["--", "#", ("/*", "*/")] 133 IDENTIFIERS = ["`"] 134 STRING_ESCAPES = ["\\"] 135 HEX_STRINGS = [("0x", ""), ("0X", "")] 136 137 KEYWORDS = { 138 **tokens.Tokenizer.KEYWORDS, 139 "BEGIN": TokenType.COMMAND, 140 "BEGIN TRANSACTION": TokenType.BEGIN, 141 "CURRENT_DATETIME": TokenType.CURRENT_DATETIME, 142 "CURRENT_TIME": TokenType.CURRENT_TIME, 143 "DECLARE": TokenType.COMMAND, 144 "GEOGRAPHY": TokenType.GEOGRAPHY, 145 "FLOAT64": TokenType.DOUBLE, 146 "INT64": TokenType.BIGINT, 147 "NOT DETERMINISTIC": TokenType.VOLATILE, 148 "UNKNOWN": TokenType.NULL, 149 } 150 KEYWORDS.pop("DIV") 151 152 class Parser(parser.Parser): 153 FUNCTIONS = { 154 **parser.Parser.FUNCTIONS, # type: ignore 155 "DATE_TRUNC": lambda args: exp.DateTrunc( 156 unit=exp.Literal.string(seq_get(args, 1).name), # type: ignore 157 this=seq_get(args, 0), 158 ), 159 "DATE_ADD": _date_add(exp.DateAdd), 160 "DATETIME_ADD": _date_add(exp.DatetimeAdd), 161 "DIV": lambda args: exp.IntDiv(this=seq_get(args, 0), expression=seq_get(args, 1)), 162 "REGEXP_CONTAINS": exp.RegexpLike.from_arg_list, 163 "TIME_ADD": _date_add(exp.TimeAdd), 164 "TIMESTAMP_ADD": _date_add(exp.TimestampAdd), 165 "DATE_SUB": _date_add(exp.DateSub), 166 "DATETIME_SUB": _date_add(exp.DatetimeSub), 167 "TIME_SUB": _date_add(exp.TimeSub), 168 "TIMESTAMP_SUB": _date_add(exp.TimestampSub), 169 "PARSE_TIMESTAMP": lambda args: exp.StrToTime( 170 this=seq_get(args, 1), format=seq_get(args, 0) 171 ), 172 } 173 174 FUNCTION_PARSERS = { 175 **parser.Parser.FUNCTION_PARSERS, # type: ignore 176 "ARRAY": lambda self: self.expression(exp.Array, expressions=[self._parse_statement()]), 177 } 178 FUNCTION_PARSERS.pop("TRIM") 179 180 NO_PAREN_FUNCTIONS = { 181 **parser.Parser.NO_PAREN_FUNCTIONS, # type: ignore 182 TokenType.CURRENT_DATETIME: exp.CurrentDatetime, 183 TokenType.CURRENT_TIME: exp.CurrentTime, 184 } 185 186 NESTED_TYPE_TOKENS = { 187 *parser.Parser.NESTED_TYPE_TOKENS, # type: ignore 188 TokenType.TABLE, 189 } 190 191 ID_VAR_TOKENS = { 192 *parser.Parser.ID_VAR_TOKENS, # type: ignore 193 TokenType.VALUES, 194 } 195 196 PROPERTY_PARSERS = { 197 **parser.Parser.PROPERTY_PARSERS, # type: ignore 198 "NOT DETERMINISTIC": lambda self: self.expression( 199 exp.VolatilityProperty, this=exp.Literal.string("VOLATILE") 200 ), 201 } 202 203 class Generator(generator.Generator): 204 TRANSFORMS = { 205 **generator.Generator.TRANSFORMS, # type: ignore 206 **transforms.REMOVE_PRECISION_PARAMETERIZED_TYPES, # type: ignore 207 exp.ArraySize: rename_func("ARRAY_LENGTH"), 208 exp.DateAdd: _date_add_sql("DATE", "ADD"), 209 exp.DateSub: _date_add_sql("DATE", "SUB"), 210 exp.DatetimeAdd: _date_add_sql("DATETIME", "ADD"), 211 exp.DatetimeSub: _date_add_sql("DATETIME", "SUB"), 212 exp.DateDiff: lambda self, e: f"DATE_DIFF({self.sql(e, 'this')}, {self.sql(e, 'expression')}, {self.sql(e.args.get('unit', 'DAY'))})", 213 exp.DateStrToDate: datestrtodate_sql, 214 exp.DateTrunc: lambda self, e: self.func("DATE_TRUNC", e.this, e.text("unit")), 215 exp.GroupConcat: rename_func("STRING_AGG"), 216 exp.ILike: no_ilike_sql, 217 exp.IntDiv: rename_func("DIV"), 218 exp.Select: transforms.preprocess( 219 [_unqualify_unnest], transforms.delegate("select_sql") 220 ), 221 exp.StrToTime: lambda self, e: f"PARSE_TIMESTAMP({self.format_time(e)}, {self.sql(e, 'this')})", 222 exp.TimeAdd: _date_add_sql("TIME", "ADD"), 223 exp.TimeSub: _date_add_sql("TIME", "SUB"), 224 exp.TimestampAdd: _date_add_sql("TIMESTAMP", "ADD"), 225 exp.TimestampSub: _date_add_sql("TIMESTAMP", "SUB"), 226 exp.TimeStrToTime: timestrtotime_sql, 227 exp.PartitionedByProperty: lambda self, e: f"PARTITION BY {self.sql(e, 'this')}", 228 exp.VariancePop: rename_func("VAR_POP"), 229 exp.Values: _derived_table_values_to_unnest, 230 exp.ReturnsProperty: _returnsproperty_sql, 231 exp.Create: _create_sql, 232 exp.Trim: lambda self, e: self.func(f"TRIM", e.this, e.expression), 233 exp.VolatilityProperty: lambda self, e: f"DETERMINISTIC" 234 if e.name == "IMMUTABLE" 235 else "NOT DETERMINISTIC", 236 exp.RegexpLike: rename_func("REGEXP_CONTAINS"), 237 } 238 239 TYPE_MAPPING = { 240 **generator.Generator.TYPE_MAPPING, # type: ignore 241 exp.DataType.Type.TINYINT: "INT64", 242 exp.DataType.Type.SMALLINT: "INT64", 243 exp.DataType.Type.INT: "INT64", 244 exp.DataType.Type.BIGINT: "INT64", 245 exp.DataType.Type.DECIMAL: "NUMERIC", 246 exp.DataType.Type.FLOAT: "FLOAT64", 247 exp.DataType.Type.DOUBLE: "FLOAT64", 248 exp.DataType.Type.BOOLEAN: "BOOL", 249 exp.DataType.Type.TEXT: "STRING", 250 exp.DataType.Type.VARCHAR: "STRING", 251 exp.DataType.Type.NVARCHAR: "STRING", 252 } 253 PROPERTIES_LOCATION = { 254 **generator.Generator.PROPERTIES_LOCATION, # type: ignore 255 exp.PartitionedByProperty: exp.Properties.Location.POST_SCHEMA, 256 } 257 258 EXPLICIT_UNION = True 259 260 def array_sql(self, expression: exp.Array) -> str: 261 first_arg = seq_get(expression.expressions, 0) 262 if isinstance(first_arg, exp.Subqueryable): 263 return f"ARRAY{self.wrap(self.sql(first_arg))}" 264 265 return inline_array_sql(self, expression) 266 267 def transaction_sql(self, *_) -> str: 268 return "BEGIN TRANSACTION" 269 270 def commit_sql(self, *_) -> str: 271 return "COMMIT TRANSACTION" 272 273 def rollback_sql(self, *_) -> str: 274 return "ROLLBACK TRANSACTION" 275 276 def in_unnest_op(self, expression: exp.Unnest) -> str: 277 return self.sql(expression) 278 279 def except_op(self, expression: exp.Except) -> str: 280 if not expression.args.get("distinct", False): 281 self.unsupported("EXCEPT without DISTINCT is not supported in BigQuery") 282 return f"EXCEPT{' DISTINCT' if expression.args.get('distinct') else ' ALL'}" 283 284 def intersect_op(self, expression: exp.Intersect) -> str: 285 if not expression.args.get("distinct", False): 286 self.unsupported("INTERSECT without DISTINCT is not supported in BigQuery") 287 return f"INTERSECT{' DISTINCT' if expression.args.get('distinct') else ' ALL'}"
Inherited Members
126 class Tokenizer(tokens.Tokenizer): 127 QUOTES = [ 128 (prefix + quote, quote) if prefix else quote 129 for quote in ["'", '"', '"""', "'''"] 130 for prefix in ["", "r", "R"] 131 ] 132 COMMENTS = ["--", "#", ("/*", "*/")] 133 IDENTIFIERS = ["`"] 134 STRING_ESCAPES = ["\\"] 135 HEX_STRINGS = [("0x", ""), ("0X", "")] 136 137 KEYWORDS = { 138 **tokens.Tokenizer.KEYWORDS, 139 "BEGIN": TokenType.COMMAND, 140 "BEGIN TRANSACTION": TokenType.BEGIN, 141 "CURRENT_DATETIME": TokenType.CURRENT_DATETIME, 142 "CURRENT_TIME": TokenType.CURRENT_TIME, 143 "DECLARE": TokenType.COMMAND, 144 "GEOGRAPHY": TokenType.GEOGRAPHY, 145 "FLOAT64": TokenType.DOUBLE, 146 "INT64": TokenType.BIGINT, 147 "NOT DETERMINISTIC": TokenType.VOLATILE, 148 "UNKNOWN": TokenType.NULL, 149 } 150 KEYWORDS.pop("DIV")
Inherited Members
152 class Parser(parser.Parser): 153 FUNCTIONS = { 154 **parser.Parser.FUNCTIONS, # type: ignore 155 "DATE_TRUNC": lambda args: exp.DateTrunc( 156 unit=exp.Literal.string(seq_get(args, 1).name), # type: ignore 157 this=seq_get(args, 0), 158 ), 159 "DATE_ADD": _date_add(exp.DateAdd), 160 "DATETIME_ADD": _date_add(exp.DatetimeAdd), 161 "DIV": lambda args: exp.IntDiv(this=seq_get(args, 0), expression=seq_get(args, 1)), 162 "REGEXP_CONTAINS": exp.RegexpLike.from_arg_list, 163 "TIME_ADD": _date_add(exp.TimeAdd), 164 "TIMESTAMP_ADD": _date_add(exp.TimestampAdd), 165 "DATE_SUB": _date_add(exp.DateSub), 166 "DATETIME_SUB": _date_add(exp.DatetimeSub), 167 "TIME_SUB": _date_add(exp.TimeSub), 168 "TIMESTAMP_SUB": _date_add(exp.TimestampSub), 169 "PARSE_TIMESTAMP": lambda args: exp.StrToTime( 170 this=seq_get(args, 1), format=seq_get(args, 0) 171 ), 172 } 173 174 FUNCTION_PARSERS = { 175 **parser.Parser.FUNCTION_PARSERS, # type: ignore 176 "ARRAY": lambda self: self.expression(exp.Array, expressions=[self._parse_statement()]), 177 } 178 FUNCTION_PARSERS.pop("TRIM") 179 180 NO_PAREN_FUNCTIONS = { 181 **parser.Parser.NO_PAREN_FUNCTIONS, # type: ignore 182 TokenType.CURRENT_DATETIME: exp.CurrentDatetime, 183 TokenType.CURRENT_TIME: exp.CurrentTime, 184 } 185 186 NESTED_TYPE_TOKENS = { 187 *parser.Parser.NESTED_TYPE_TOKENS, # type: ignore 188 TokenType.TABLE, 189 } 190 191 ID_VAR_TOKENS = { 192 *parser.Parser.ID_VAR_TOKENS, # type: ignore 193 TokenType.VALUES, 194 } 195 196 PROPERTY_PARSERS = { 197 **parser.Parser.PROPERTY_PARSERS, # type: ignore 198 "NOT DETERMINISTIC": lambda self: self.expression( 199 exp.VolatilityProperty, this=exp.Literal.string("VOLATILE") 200 ), 201 }
Parser consumes a list of tokens produced by the sqlglot.tokens.Tokenizer
and produces
a parsed syntax tree.
Arguments:
- error_level: the desired error level. Default: ErrorLevel.RAISE
- error_message_context: determines the amount of context to capture from a query string when displaying the error message (in number of characters). Default: 50.
- index_offset: Index offset for arrays eg ARRAY[0] vs ARRAY[1] as the head of a list. Default: 0
- alias_post_tablesample: If the table alias comes after tablesample. Default: False
- 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
- null_ordering: Indicates the default null ordering method to use if not explicitly set. Options are "nulls_are_small", "nulls_are_large", "nulls_are_last". Default: "nulls_are_small"
Inherited Members
203 class Generator(generator.Generator): 204 TRANSFORMS = { 205 **generator.Generator.TRANSFORMS, # type: ignore 206 **transforms.REMOVE_PRECISION_PARAMETERIZED_TYPES, # type: ignore 207 exp.ArraySize: rename_func("ARRAY_LENGTH"), 208 exp.DateAdd: _date_add_sql("DATE", "ADD"), 209 exp.DateSub: _date_add_sql("DATE", "SUB"), 210 exp.DatetimeAdd: _date_add_sql("DATETIME", "ADD"), 211 exp.DatetimeSub: _date_add_sql("DATETIME", "SUB"), 212 exp.DateDiff: lambda self, e: f"DATE_DIFF({self.sql(e, 'this')}, {self.sql(e, 'expression')}, {self.sql(e.args.get('unit', 'DAY'))})", 213 exp.DateStrToDate: datestrtodate_sql, 214 exp.DateTrunc: lambda self, e: self.func("DATE_TRUNC", e.this, e.text("unit")), 215 exp.GroupConcat: rename_func("STRING_AGG"), 216 exp.ILike: no_ilike_sql, 217 exp.IntDiv: rename_func("DIV"), 218 exp.Select: transforms.preprocess( 219 [_unqualify_unnest], transforms.delegate("select_sql") 220 ), 221 exp.StrToTime: lambda self, e: f"PARSE_TIMESTAMP({self.format_time(e)}, {self.sql(e, 'this')})", 222 exp.TimeAdd: _date_add_sql("TIME", "ADD"), 223 exp.TimeSub: _date_add_sql("TIME", "SUB"), 224 exp.TimestampAdd: _date_add_sql("TIMESTAMP", "ADD"), 225 exp.TimestampSub: _date_add_sql("TIMESTAMP", "SUB"), 226 exp.TimeStrToTime: timestrtotime_sql, 227 exp.PartitionedByProperty: lambda self, e: f"PARTITION BY {self.sql(e, 'this')}", 228 exp.VariancePop: rename_func("VAR_POP"), 229 exp.Values: _derived_table_values_to_unnest, 230 exp.ReturnsProperty: _returnsproperty_sql, 231 exp.Create: _create_sql, 232 exp.Trim: lambda self, e: self.func(f"TRIM", e.this, e.expression), 233 exp.VolatilityProperty: lambda self, e: f"DETERMINISTIC" 234 if e.name == "IMMUTABLE" 235 else "NOT DETERMINISTIC", 236 exp.RegexpLike: rename_func("REGEXP_CONTAINS"), 237 } 238 239 TYPE_MAPPING = { 240 **generator.Generator.TYPE_MAPPING, # type: ignore 241 exp.DataType.Type.TINYINT: "INT64", 242 exp.DataType.Type.SMALLINT: "INT64", 243 exp.DataType.Type.INT: "INT64", 244 exp.DataType.Type.BIGINT: "INT64", 245 exp.DataType.Type.DECIMAL: "NUMERIC", 246 exp.DataType.Type.FLOAT: "FLOAT64", 247 exp.DataType.Type.DOUBLE: "FLOAT64", 248 exp.DataType.Type.BOOLEAN: "BOOL", 249 exp.DataType.Type.TEXT: "STRING", 250 exp.DataType.Type.VARCHAR: "STRING", 251 exp.DataType.Type.NVARCHAR: "STRING", 252 } 253 PROPERTIES_LOCATION = { 254 **generator.Generator.PROPERTIES_LOCATION, # type: ignore 255 exp.PartitionedByProperty: exp.Properties.Location.POST_SCHEMA, 256 } 257 258 EXPLICIT_UNION = True 259 260 def array_sql(self, expression: exp.Array) -> str: 261 first_arg = seq_get(expression.expressions, 0) 262 if isinstance(first_arg, exp.Subqueryable): 263 return f"ARRAY{self.wrap(self.sql(first_arg))}" 264 265 return inline_array_sql(self, expression) 266 267 def transaction_sql(self, *_) -> str: 268 return "BEGIN TRANSACTION" 269 270 def commit_sql(self, *_) -> str: 271 return "COMMIT TRANSACTION" 272 273 def rollback_sql(self, *_) -> str: 274 return "ROLLBACK TRANSACTION" 275 276 def in_unnest_op(self, expression: exp.Unnest) -> str: 277 return self.sql(expression) 278 279 def except_op(self, expression: exp.Except) -> str: 280 if not expression.args.get("distinct", False): 281 self.unsupported("EXCEPT without DISTINCT is not supported in BigQuery") 282 return f"EXCEPT{' DISTINCT' if expression.args.get('distinct') else ' ALL'}" 283 284 def intersect_op(self, expression: exp.Intersect) -> str: 285 if not expression.args.get("distinct", False): 286 self.unsupported("INTERSECT without DISTINCT is not supported in BigQuery") 287 return f"INTERSECT{' DISTINCT' if expression.args.get('distinct') else ' ALL'}"
Generator interprets the given syntax tree and produces a SQL string as an output.
Arguments:
- time_mapping (dict): the dictionary of custom time mappings in which the key represents a python time format and the output the target time format
- time_trie (trie): a trie of the time_mapping keys
- pretty (bool): if set to True the returned string will be formatted. Default: False.
- quote_start (str): specifies which starting character to use to delimit quotes. Default: '.
- quote_end (str): specifies which ending character to use to delimit quotes. Default: '.
- identifier_start (str): specifies which starting character to use to delimit identifiers. Default: ".
- identifier_end (str): specifies which ending character to use to delimit identifiers. Default: ".
- identify (bool): if set to True all identifiers will be delimited by the corresponding character.
- normalize (bool): if set to True all identifiers will lower cased
- string_escape (str): specifies a string escape character. Default: '.
- identifier_escape (str): specifies an identifier escape character. Default: ".
- pad (int): determines padding in a formatted string. Default: 2.
- indent (int): determines the size of indentation in a formatted string. Default: 4.
- unnest_column_only (bool): if true unnest table aliases are considered only as column aliases
- normalize_functions (str): normalize function names, "upper", "lower", or None Default: "upper"
- alias_post_tablesample (bool): if the table alias comes after tablesample Default: False
- unsupported_level (ErrorLevel): determines the generator's behavior when it encounters unsupported expressions. Default ErrorLevel.WARN.
- null_ordering (str): Indicates the default null ordering method to use if not explicitly set. Options are "nulls_are_small", "nulls_are_large", "nulls_are_last". Default: "nulls_are_small"
- max_unsupported (int): 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 (bool): if the the comma is leading or trailing in select statements 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
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
- columndef_sql
- columnconstraint_sql
- autoincrementcolumnconstraint_sql
- generatedasidentitycolumnconstraint_sql
- notnullcolumnconstraint_sql
- primarykeycolumnconstraint_sql
- uniquecolumnconstraint_sql
- create_sql
- describe_sql
- prepend_ctes
- with_sql
- cte_sql
- tablealias_sql
- bitstring_sql
- hexstring_sql
- datatype_sql
- directory_sql
- delete_sql
- drop_sql
- except_sql
- fetch_sql
- filter_sql
- hint_sql
- index_sql
- identifier_sql
- national_sql
- partition_sql
- properties_sql
- root_properties
- properties
- with_properties
- locate_properties
- property_sql
- likeproperty_sql
- fallbackproperty_sql
- journalproperty_sql
- freespaceproperty_sql
- afterjournalproperty_sql
- checksumproperty_sql
- mergeblockratioproperty_sql
- datablocksizeproperty_sql
- blockcompressionproperty_sql
- isolatedloadingproperty_sql
- lockingproperty_sql
- insert_sql
- intersect_sql
- introducer_sql
- pseudotype_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
- lock_sql
- literal_sql
- loaddata_sql
- null_sql
- boolean_sql
- order_sql
- cluster_sql
- distribute_sql
- sort_sql
- ordered_sql
- matchrecognize_sql
- query_modifiers
- select_sql
- schema_sql
- star_sql
- structkwarg_sql
- parameter_sql
- sessionparameter_sql
- placeholder_sql
- subquery_sql
- qualify_sql
- union_sql
- union_op
- unnest_sql
- where_sql
- window_sql
- partition_by_sql
- window_spec_sql
- withingroup_sql
- between_sql
- bracket_sql
- all_sql
- any_sql
- exists_sql
- case_sql
- constraint_sql
- extract_sql
- trim_sql
- concat_sql
- check_sql
- foreignkey_sql
- primarykey_sql
- unique_sql
- if_sql
- in_sql
- 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
- altercolumn_sql
- renametable_sql
- altertable_sql
- droppartition_sql
- addconstraint_sql
- distinct_sql
- ignorenulls_sql
- respectnulls_sql
- intdiv_sql
- dpipe_sql
- div_sql
- distance_sql
- dot_sql
- eq_sql
- escape_sql
- glob_sql
- gt_sql
- gte_sql
- ilike_sql
- is_sql
- like_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