sqlglot.dialects.postgres
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 arrow_json_extract_scalar_sql, 9 arrow_json_extract_sql, 10 datestrtodate_sql, 11 format_time_lambda, 12 max_or_greatest, 13 min_or_least, 14 no_paren_current_date_sql, 15 no_pivot_sql, 16 no_tablesample_sql, 17 no_trycast_sql, 18 rename_func, 19 str_position_sql, 20 timestamptrunc_sql, 21 timestrtotime_sql, 22 trim_sql, 23 ts_or_ds_to_date_sql, 24) 25from sqlglot.helper import seq_get 26from sqlglot.parser import binary_range_parser 27from sqlglot.tokens import TokenType 28 29DATE_DIFF_FACTOR = { 30 "MICROSECOND": " * 1000000", 31 "MILLISECOND": " * 1000", 32 "SECOND": "", 33 "MINUTE": " / 60", 34 "HOUR": " / 3600", 35 "DAY": " / 86400", 36} 37 38 39def _date_add_sql(kind: str) -> t.Callable[[generator.Generator, exp.DateAdd | exp.DateSub], str]: 40 def func(self: generator.Generator, expression: exp.DateAdd | exp.DateSub) -> str: 41 from sqlglot.optimizer.simplify import simplify 42 43 this = self.sql(expression, "this") 44 unit = expression.args.get("unit") 45 expression = simplify(expression.args["expression"]) 46 47 if not isinstance(expression, exp.Literal): 48 self.unsupported("Cannot add non literal") 49 50 expression = expression.copy() 51 expression.args["is_string"] = True 52 return f"{this} {kind} {self.sql(exp.Interval(this=expression, unit=unit))}" 53 54 return func 55 56 57def _date_diff_sql(self: generator.Generator, expression: exp.DateDiff) -> str: 58 unit = expression.text("unit").upper() 59 factor = DATE_DIFF_FACTOR.get(unit) 60 61 end = f"CAST({expression.this} AS TIMESTAMP)" 62 start = f"CAST({expression.expression} AS TIMESTAMP)" 63 64 if factor is not None: 65 return f"CAST(EXTRACT(epoch FROM {end} - {start}){factor} AS BIGINT)" 66 67 age = f"AGE({end}, {start})" 68 69 if unit == "WEEK": 70 unit = f"EXTRACT(year FROM {age}) * 48 + EXTRACT(month FROM {age}) * 4 + EXTRACT(day FROM {age}) / 7" 71 elif unit == "MONTH": 72 unit = f"EXTRACT(year FROM {age}) * 12 + EXTRACT(month FROM {age})" 73 elif unit == "QUARTER": 74 unit = f"EXTRACT(year FROM {age}) * 4 + EXTRACT(month FROM {age}) / 3" 75 elif unit == "YEAR": 76 unit = f"EXTRACT(year FROM {age})" 77 else: 78 unit = age 79 80 return f"CAST({unit} AS BIGINT)" 81 82 83def _substring_sql(self: generator.Generator, expression: exp.Substring) -> str: 84 this = self.sql(expression, "this") 85 start = self.sql(expression, "start") 86 length = self.sql(expression, "length") 87 88 from_part = f" FROM {start}" if start else "" 89 for_part = f" FOR {length}" if length else "" 90 91 return f"SUBSTRING({this}{from_part}{for_part})" 92 93 94def _string_agg_sql(self: generator.Generator, expression: exp.GroupConcat) -> str: 95 expression = expression.copy() 96 separator = expression.args.get("separator") or exp.Literal.string(",") 97 98 order = "" 99 this = expression.this 100 if isinstance(this, exp.Order): 101 if this.this: 102 this = this.this.pop() 103 order = self.sql(expression.this) # Order has a leading space 104 105 return f"STRING_AGG({self.format_args(this, separator)}{order})" 106 107 108def _datatype_sql(self: generator.Generator, expression: exp.DataType) -> str: 109 if expression.is_type("array"): 110 return f"{self.expressions(expression, flat=True)}[]" 111 return self.datatype_sql(expression) 112 113 114def _auto_increment_to_serial(expression: exp.Expression) -> exp.Expression: 115 auto = expression.find(exp.AutoIncrementColumnConstraint) 116 117 if auto: 118 expression = expression.copy() 119 expression.args["constraints"].remove(auto.parent) 120 kind = expression.args["kind"] 121 122 if kind.this == exp.DataType.Type.INT: 123 kind.replace(exp.DataType(this=exp.DataType.Type.SERIAL)) 124 elif kind.this == exp.DataType.Type.SMALLINT: 125 kind.replace(exp.DataType(this=exp.DataType.Type.SMALLSERIAL)) 126 elif kind.this == exp.DataType.Type.BIGINT: 127 kind.replace(exp.DataType(this=exp.DataType.Type.BIGSERIAL)) 128 129 return expression 130 131 132def _serial_to_generated(expression: exp.Expression) -> exp.Expression: 133 kind = expression.args["kind"] 134 135 if kind.this == exp.DataType.Type.SERIAL: 136 data_type = exp.DataType(this=exp.DataType.Type.INT) 137 elif kind.this == exp.DataType.Type.SMALLSERIAL: 138 data_type = exp.DataType(this=exp.DataType.Type.SMALLINT) 139 elif kind.this == exp.DataType.Type.BIGSERIAL: 140 data_type = exp.DataType(this=exp.DataType.Type.BIGINT) 141 else: 142 data_type = None 143 144 if data_type: 145 expression = expression.copy() 146 expression.args["kind"].replace(data_type) 147 constraints = expression.args["constraints"] 148 generated = exp.ColumnConstraint(kind=exp.GeneratedAsIdentityColumnConstraint(this=False)) 149 notnull = exp.ColumnConstraint(kind=exp.NotNullColumnConstraint()) 150 151 if notnull not in constraints: 152 constraints.insert(0, notnull) 153 if generated not in constraints: 154 constraints.insert(0, generated) 155 156 return expression 157 158 159def _generate_series(args: t.List) -> exp.Expression: 160 # The goal is to convert step values like '1 day' or INTERVAL '1 day' into INTERVAL '1' day 161 step = seq_get(args, 2) 162 163 if step is None: 164 # Postgres allows calls with just two arguments -- the "step" argument defaults to 1 165 return exp.GenerateSeries.from_arg_list(args) 166 167 if step.is_string: 168 args[2] = exp.to_interval(step.this) 169 elif isinstance(step, exp.Interval) and not step.args.get("unit"): 170 args[2] = exp.to_interval(step.this.this) 171 172 return exp.GenerateSeries.from_arg_list(args) 173 174 175def _to_timestamp(args: t.List) -> exp.Expression: 176 # TO_TIMESTAMP accepts either a single double argument or (text, text) 177 if len(args) == 1: 178 # https://www.postgresql.org/docs/current/functions-datetime.html#FUNCTIONS-DATETIME-TABLE 179 return exp.UnixToTime.from_arg_list(args) 180 181 # https://www.postgresql.org/docs/current/functions-formatting.html 182 return format_time_lambda(exp.StrToTime, "postgres")(args) 183 184 185class Postgres(Dialect): 186 INDEX_OFFSET = 1 187 NULL_ORDERING = "nulls_are_large" 188 TIME_FORMAT = "'YYYY-MM-DD HH24:MI:SS'" 189 TIME_MAPPING = { 190 "AM": "%p", 191 "PM": "%p", 192 "D": "%u", # 1-based day of week 193 "DD": "%d", # day of month 194 "DDD": "%j", # zero padded day of year 195 "FMDD": "%-d", # - is no leading zero for Python; same for FM in postgres 196 "FMDDD": "%-j", # day of year 197 "FMHH12": "%-I", # 9 198 "FMHH24": "%-H", # 9 199 "FMMI": "%-M", # Minute 200 "FMMM": "%-m", # 1 201 "FMSS": "%-S", # Second 202 "HH12": "%I", # 09 203 "HH24": "%H", # 09 204 "MI": "%M", # zero padded minute 205 "MM": "%m", # 01 206 "OF": "%z", # utc offset 207 "SS": "%S", # zero padded second 208 "TMDay": "%A", # TM is locale dependent 209 "TMDy": "%a", 210 "TMMon": "%b", # Sep 211 "TMMonth": "%B", # September 212 "TZ": "%Z", # uppercase timezone name 213 "US": "%f", # zero padded microsecond 214 "WW": "%U", # 1-based week of year 215 "YY": "%y", # 15 216 "YYYY": "%Y", # 2015 217 } 218 219 class Tokenizer(tokens.Tokenizer): 220 QUOTES = ["'", "$$"] 221 222 BIT_STRINGS = [("b'", "'"), ("B'", "'")] 223 HEX_STRINGS = [("x'", "'"), ("X'", "'")] 224 BYTE_STRINGS = [("e'", "'"), ("E'", "'")] 225 226 KEYWORDS = { 227 **tokens.Tokenizer.KEYWORDS, 228 "~~": TokenType.LIKE, 229 "~~*": TokenType.ILIKE, 230 "~*": TokenType.IRLIKE, 231 "~": TokenType.RLIKE, 232 "@>": TokenType.AT_GT, 233 "<@": TokenType.LT_AT, 234 "BEGIN": TokenType.COMMAND, 235 "BEGIN TRANSACTION": TokenType.BEGIN, 236 "BIGSERIAL": TokenType.BIGSERIAL, 237 "CHARACTER VARYING": TokenType.VARCHAR, 238 "DECLARE": TokenType.COMMAND, 239 "DO": TokenType.COMMAND, 240 "HSTORE": TokenType.HSTORE, 241 "JSONB": TokenType.JSONB, 242 "REFRESH": TokenType.COMMAND, 243 "REINDEX": TokenType.COMMAND, 244 "RESET": TokenType.COMMAND, 245 "REVOKE": TokenType.COMMAND, 246 "SERIAL": TokenType.SERIAL, 247 "SMALLSERIAL": TokenType.SMALLSERIAL, 248 "TEMP": TokenType.TEMPORARY, 249 "CSTRING": TokenType.PSEUDO_TYPE, 250 } 251 252 SINGLE_TOKENS = { 253 **tokens.Tokenizer.SINGLE_TOKENS, 254 "$": TokenType.PARAMETER, 255 } 256 257 VAR_SINGLE_TOKENS = {"$"} 258 259 class Parser(parser.Parser): 260 STRICT_CAST = False 261 CONCAT_NULL_OUTPUTS_STRING = True 262 263 FUNCTIONS = { 264 **parser.Parser.FUNCTIONS, 265 "DATE_TRUNC": lambda args: exp.TimestampTrunc( 266 this=seq_get(args, 1), unit=seq_get(args, 0) 267 ), 268 "GENERATE_SERIES": _generate_series, 269 "NOW": exp.CurrentTimestamp.from_arg_list, 270 "TO_CHAR": format_time_lambda(exp.TimeToStr, "postgres"), 271 "TO_TIMESTAMP": _to_timestamp, 272 "UNNEST": exp.Explode.from_arg_list, 273 } 274 275 FUNCTION_PARSERS = { 276 **parser.Parser.FUNCTION_PARSERS, 277 "DATE_PART": lambda self: self._parse_date_part(), 278 } 279 280 BITWISE = { 281 **parser.Parser.BITWISE, 282 TokenType.HASH: exp.BitwiseXor, 283 } 284 285 EXPONENT = { 286 TokenType.CARET: exp.Pow, 287 } 288 289 RANGE_PARSERS = { 290 **parser.Parser.RANGE_PARSERS, 291 TokenType.DAMP: binary_range_parser(exp.ArrayOverlaps), 292 TokenType.AT_GT: binary_range_parser(exp.ArrayContains), 293 TokenType.LT_AT: binary_range_parser(exp.ArrayContained), 294 } 295 296 def _parse_factor(self) -> t.Optional[exp.Expression]: 297 return self._parse_tokens(self._parse_exponent, self.FACTOR) 298 299 def _parse_exponent(self) -> t.Optional[exp.Expression]: 300 return self._parse_tokens(self._parse_unary, self.EXPONENT) 301 302 def _parse_date_part(self) -> exp.Expression: 303 part = self._parse_type() 304 self._match(TokenType.COMMA) 305 value = self._parse_bitwise() 306 307 if part and part.is_string: 308 part = exp.var(part.name) 309 310 return self.expression(exp.Extract, this=part, expression=value) 311 312 class Generator(generator.Generator): 313 SINGLE_STRING_INTERVAL = True 314 LOCKING_READS_SUPPORTED = True 315 JOIN_HINTS = False 316 TABLE_HINTS = False 317 PARAMETER_TOKEN = "$" 318 319 TYPE_MAPPING = { 320 **generator.Generator.TYPE_MAPPING, 321 exp.DataType.Type.TINYINT: "SMALLINT", 322 exp.DataType.Type.FLOAT: "REAL", 323 exp.DataType.Type.DOUBLE: "DOUBLE PRECISION", 324 exp.DataType.Type.BINARY: "BYTEA", 325 exp.DataType.Type.VARBINARY: "BYTEA", 326 exp.DataType.Type.DATETIME: "TIMESTAMP", 327 } 328 329 TRANSFORMS = { 330 **generator.Generator.TRANSFORMS, 331 exp.BitwiseXor: lambda self, e: self.binary(e, "#"), 332 exp.ColumnDef: transforms.preprocess([_auto_increment_to_serial, _serial_to_generated]), 333 exp.Explode: rename_func("UNNEST"), 334 exp.JSONExtract: arrow_json_extract_sql, 335 exp.JSONExtractScalar: arrow_json_extract_scalar_sql, 336 exp.JSONBExtract: lambda self, e: self.binary(e, "#>"), 337 exp.JSONBExtractScalar: lambda self, e: self.binary(e, "#>>"), 338 exp.JSONBContains: lambda self, e: self.binary(e, "?"), 339 exp.Pow: lambda self, e: self.binary(e, "^"), 340 exp.CurrentDate: no_paren_current_date_sql, 341 exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP", 342 exp.DateAdd: _date_add_sql("+"), 343 exp.DateStrToDate: datestrtodate_sql, 344 exp.DateSub: _date_add_sql("-"), 345 exp.DateDiff: _date_diff_sql, 346 exp.LogicalOr: rename_func("BOOL_OR"), 347 exp.LogicalAnd: rename_func("BOOL_AND"), 348 exp.Max: max_or_greatest, 349 exp.Min: min_or_least, 350 exp.ArrayOverlaps: lambda self, e: self.binary(e, "&&"), 351 exp.ArrayContains: lambda self, e: self.binary(e, "@>"), 352 exp.ArrayContained: lambda self, e: self.binary(e, "<@"), 353 exp.Merge: transforms.preprocess([transforms.remove_target_from_merge]), 354 exp.Pivot: no_pivot_sql, 355 exp.RegexpLike: lambda self, e: self.binary(e, "~"), 356 exp.RegexpILike: lambda self, e: self.binary(e, "~*"), 357 exp.StrPosition: str_position_sql, 358 exp.StrToTime: lambda self, e: f"TO_TIMESTAMP({self.sql(e, 'this')}, {self.format_time(e)})", 359 exp.Substring: _substring_sql, 360 exp.TimestampTrunc: timestamptrunc_sql, 361 exp.TimeStrToTime: timestrtotime_sql, 362 exp.TimeToStr: lambda self, e: f"TO_CHAR({self.sql(e, 'this')}, {self.format_time(e)})", 363 exp.TableSample: no_tablesample_sql, 364 exp.ToChar: lambda self, e: self.function_fallback_sql(e), 365 exp.Trim: trim_sql, 366 exp.TryCast: no_trycast_sql, 367 exp.TsOrDsToDate: ts_or_ds_to_date_sql("postgres"), 368 exp.UnixToTime: lambda self, e: f"TO_TIMESTAMP({self.sql(e, 'this')})", 369 exp.DataType: _datatype_sql, 370 exp.GroupConcat: _string_agg_sql, 371 exp.Array: lambda self, e: f"{self.normalize_func('ARRAY')}({self.sql(e.expressions[0])})" 372 if isinstance(seq_get(e.expressions, 0), exp.Select) 373 else f"{self.normalize_func('ARRAY')}[{self.expressions(e, flat=True)}]", 374 } 375 376 PROPERTIES_LOCATION = { 377 **generator.Generator.PROPERTIES_LOCATION, 378 exp.TransientProperty: exp.Properties.Location.UNSUPPORTED, 379 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 380 }
186class Postgres(Dialect): 187 INDEX_OFFSET = 1 188 NULL_ORDERING = "nulls_are_large" 189 TIME_FORMAT = "'YYYY-MM-DD HH24:MI:SS'" 190 TIME_MAPPING = { 191 "AM": "%p", 192 "PM": "%p", 193 "D": "%u", # 1-based day of week 194 "DD": "%d", # day of month 195 "DDD": "%j", # zero padded day of year 196 "FMDD": "%-d", # - is no leading zero for Python; same for FM in postgres 197 "FMDDD": "%-j", # day of year 198 "FMHH12": "%-I", # 9 199 "FMHH24": "%-H", # 9 200 "FMMI": "%-M", # Minute 201 "FMMM": "%-m", # 1 202 "FMSS": "%-S", # Second 203 "HH12": "%I", # 09 204 "HH24": "%H", # 09 205 "MI": "%M", # zero padded minute 206 "MM": "%m", # 01 207 "OF": "%z", # utc offset 208 "SS": "%S", # zero padded second 209 "TMDay": "%A", # TM is locale dependent 210 "TMDy": "%a", 211 "TMMon": "%b", # Sep 212 "TMMonth": "%B", # September 213 "TZ": "%Z", # uppercase timezone name 214 "US": "%f", # zero padded microsecond 215 "WW": "%U", # 1-based week of year 216 "YY": "%y", # 15 217 "YYYY": "%Y", # 2015 218 } 219 220 class Tokenizer(tokens.Tokenizer): 221 QUOTES = ["'", "$$"] 222 223 BIT_STRINGS = [("b'", "'"), ("B'", "'")] 224 HEX_STRINGS = [("x'", "'"), ("X'", "'")] 225 BYTE_STRINGS = [("e'", "'"), ("E'", "'")] 226 227 KEYWORDS = { 228 **tokens.Tokenizer.KEYWORDS, 229 "~~": TokenType.LIKE, 230 "~~*": TokenType.ILIKE, 231 "~*": TokenType.IRLIKE, 232 "~": TokenType.RLIKE, 233 "@>": TokenType.AT_GT, 234 "<@": TokenType.LT_AT, 235 "BEGIN": TokenType.COMMAND, 236 "BEGIN TRANSACTION": TokenType.BEGIN, 237 "BIGSERIAL": TokenType.BIGSERIAL, 238 "CHARACTER VARYING": TokenType.VARCHAR, 239 "DECLARE": TokenType.COMMAND, 240 "DO": TokenType.COMMAND, 241 "HSTORE": TokenType.HSTORE, 242 "JSONB": TokenType.JSONB, 243 "REFRESH": TokenType.COMMAND, 244 "REINDEX": TokenType.COMMAND, 245 "RESET": TokenType.COMMAND, 246 "REVOKE": TokenType.COMMAND, 247 "SERIAL": TokenType.SERIAL, 248 "SMALLSERIAL": TokenType.SMALLSERIAL, 249 "TEMP": TokenType.TEMPORARY, 250 "CSTRING": TokenType.PSEUDO_TYPE, 251 } 252 253 SINGLE_TOKENS = { 254 **tokens.Tokenizer.SINGLE_TOKENS, 255 "$": TokenType.PARAMETER, 256 } 257 258 VAR_SINGLE_TOKENS = {"$"} 259 260 class Parser(parser.Parser): 261 STRICT_CAST = False 262 CONCAT_NULL_OUTPUTS_STRING = True 263 264 FUNCTIONS = { 265 **parser.Parser.FUNCTIONS, 266 "DATE_TRUNC": lambda args: exp.TimestampTrunc( 267 this=seq_get(args, 1), unit=seq_get(args, 0) 268 ), 269 "GENERATE_SERIES": _generate_series, 270 "NOW": exp.CurrentTimestamp.from_arg_list, 271 "TO_CHAR": format_time_lambda(exp.TimeToStr, "postgres"), 272 "TO_TIMESTAMP": _to_timestamp, 273 "UNNEST": exp.Explode.from_arg_list, 274 } 275 276 FUNCTION_PARSERS = { 277 **parser.Parser.FUNCTION_PARSERS, 278 "DATE_PART": lambda self: self._parse_date_part(), 279 } 280 281 BITWISE = { 282 **parser.Parser.BITWISE, 283 TokenType.HASH: exp.BitwiseXor, 284 } 285 286 EXPONENT = { 287 TokenType.CARET: exp.Pow, 288 } 289 290 RANGE_PARSERS = { 291 **parser.Parser.RANGE_PARSERS, 292 TokenType.DAMP: binary_range_parser(exp.ArrayOverlaps), 293 TokenType.AT_GT: binary_range_parser(exp.ArrayContains), 294 TokenType.LT_AT: binary_range_parser(exp.ArrayContained), 295 } 296 297 def _parse_factor(self) -> t.Optional[exp.Expression]: 298 return self._parse_tokens(self._parse_exponent, self.FACTOR) 299 300 def _parse_exponent(self) -> t.Optional[exp.Expression]: 301 return self._parse_tokens(self._parse_unary, self.EXPONENT) 302 303 def _parse_date_part(self) -> exp.Expression: 304 part = self._parse_type() 305 self._match(TokenType.COMMA) 306 value = self._parse_bitwise() 307 308 if part and part.is_string: 309 part = exp.var(part.name) 310 311 return self.expression(exp.Extract, this=part, expression=value) 312 313 class Generator(generator.Generator): 314 SINGLE_STRING_INTERVAL = True 315 LOCKING_READS_SUPPORTED = True 316 JOIN_HINTS = False 317 TABLE_HINTS = False 318 PARAMETER_TOKEN = "$" 319 320 TYPE_MAPPING = { 321 **generator.Generator.TYPE_MAPPING, 322 exp.DataType.Type.TINYINT: "SMALLINT", 323 exp.DataType.Type.FLOAT: "REAL", 324 exp.DataType.Type.DOUBLE: "DOUBLE PRECISION", 325 exp.DataType.Type.BINARY: "BYTEA", 326 exp.DataType.Type.VARBINARY: "BYTEA", 327 exp.DataType.Type.DATETIME: "TIMESTAMP", 328 } 329 330 TRANSFORMS = { 331 **generator.Generator.TRANSFORMS, 332 exp.BitwiseXor: lambda self, e: self.binary(e, "#"), 333 exp.ColumnDef: transforms.preprocess([_auto_increment_to_serial, _serial_to_generated]), 334 exp.Explode: rename_func("UNNEST"), 335 exp.JSONExtract: arrow_json_extract_sql, 336 exp.JSONExtractScalar: arrow_json_extract_scalar_sql, 337 exp.JSONBExtract: lambda self, e: self.binary(e, "#>"), 338 exp.JSONBExtractScalar: lambda self, e: self.binary(e, "#>>"), 339 exp.JSONBContains: lambda self, e: self.binary(e, "?"), 340 exp.Pow: lambda self, e: self.binary(e, "^"), 341 exp.CurrentDate: no_paren_current_date_sql, 342 exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP", 343 exp.DateAdd: _date_add_sql("+"), 344 exp.DateStrToDate: datestrtodate_sql, 345 exp.DateSub: _date_add_sql("-"), 346 exp.DateDiff: _date_diff_sql, 347 exp.LogicalOr: rename_func("BOOL_OR"), 348 exp.LogicalAnd: rename_func("BOOL_AND"), 349 exp.Max: max_or_greatest, 350 exp.Min: min_or_least, 351 exp.ArrayOverlaps: lambda self, e: self.binary(e, "&&"), 352 exp.ArrayContains: lambda self, e: self.binary(e, "@>"), 353 exp.ArrayContained: lambda self, e: self.binary(e, "<@"), 354 exp.Merge: transforms.preprocess([transforms.remove_target_from_merge]), 355 exp.Pivot: no_pivot_sql, 356 exp.RegexpLike: lambda self, e: self.binary(e, "~"), 357 exp.RegexpILike: lambda self, e: self.binary(e, "~*"), 358 exp.StrPosition: str_position_sql, 359 exp.StrToTime: lambda self, e: f"TO_TIMESTAMP({self.sql(e, 'this')}, {self.format_time(e)})", 360 exp.Substring: _substring_sql, 361 exp.TimestampTrunc: timestamptrunc_sql, 362 exp.TimeStrToTime: timestrtotime_sql, 363 exp.TimeToStr: lambda self, e: f"TO_CHAR({self.sql(e, 'this')}, {self.format_time(e)})", 364 exp.TableSample: no_tablesample_sql, 365 exp.ToChar: lambda self, e: self.function_fallback_sql(e), 366 exp.Trim: trim_sql, 367 exp.TryCast: no_trycast_sql, 368 exp.TsOrDsToDate: ts_or_ds_to_date_sql("postgres"), 369 exp.UnixToTime: lambda self, e: f"TO_TIMESTAMP({self.sql(e, 'this')})", 370 exp.DataType: _datatype_sql, 371 exp.GroupConcat: _string_agg_sql, 372 exp.Array: lambda self, e: f"{self.normalize_func('ARRAY')}({self.sql(e.expressions[0])})" 373 if isinstance(seq_get(e.expressions, 0), exp.Select) 374 else f"{self.normalize_func('ARRAY')}[{self.expressions(e, flat=True)}]", 375 } 376 377 PROPERTIES_LOCATION = { 378 **generator.Generator.PROPERTIES_LOCATION, 379 exp.TransientProperty: exp.Properties.Location.UNSUPPORTED, 380 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 381 }
220 class Tokenizer(tokens.Tokenizer): 221 QUOTES = ["'", "$$"] 222 223 BIT_STRINGS = [("b'", "'"), ("B'", "'")] 224 HEX_STRINGS = [("x'", "'"), ("X'", "'")] 225 BYTE_STRINGS = [("e'", "'"), ("E'", "'")] 226 227 KEYWORDS = { 228 **tokens.Tokenizer.KEYWORDS, 229 "~~": TokenType.LIKE, 230 "~~*": TokenType.ILIKE, 231 "~*": TokenType.IRLIKE, 232 "~": TokenType.RLIKE, 233 "@>": TokenType.AT_GT, 234 "<@": TokenType.LT_AT, 235 "BEGIN": TokenType.COMMAND, 236 "BEGIN TRANSACTION": TokenType.BEGIN, 237 "BIGSERIAL": TokenType.BIGSERIAL, 238 "CHARACTER VARYING": TokenType.VARCHAR, 239 "DECLARE": TokenType.COMMAND, 240 "DO": TokenType.COMMAND, 241 "HSTORE": TokenType.HSTORE, 242 "JSONB": TokenType.JSONB, 243 "REFRESH": TokenType.COMMAND, 244 "REINDEX": TokenType.COMMAND, 245 "RESET": TokenType.COMMAND, 246 "REVOKE": TokenType.COMMAND, 247 "SERIAL": TokenType.SERIAL, 248 "SMALLSERIAL": TokenType.SMALLSERIAL, 249 "TEMP": TokenType.TEMPORARY, 250 "CSTRING": TokenType.PSEUDO_TYPE, 251 } 252 253 SINGLE_TOKENS = { 254 **tokens.Tokenizer.SINGLE_TOKENS, 255 "$": TokenType.PARAMETER, 256 } 257 258 VAR_SINGLE_TOKENS = {"$"}
Inherited Members
260 class Parser(parser.Parser): 261 STRICT_CAST = False 262 CONCAT_NULL_OUTPUTS_STRING = True 263 264 FUNCTIONS = { 265 **parser.Parser.FUNCTIONS, 266 "DATE_TRUNC": lambda args: exp.TimestampTrunc( 267 this=seq_get(args, 1), unit=seq_get(args, 0) 268 ), 269 "GENERATE_SERIES": _generate_series, 270 "NOW": exp.CurrentTimestamp.from_arg_list, 271 "TO_CHAR": format_time_lambda(exp.TimeToStr, "postgres"), 272 "TO_TIMESTAMP": _to_timestamp, 273 "UNNEST": exp.Explode.from_arg_list, 274 } 275 276 FUNCTION_PARSERS = { 277 **parser.Parser.FUNCTION_PARSERS, 278 "DATE_PART": lambda self: self._parse_date_part(), 279 } 280 281 BITWISE = { 282 **parser.Parser.BITWISE, 283 TokenType.HASH: exp.BitwiseXor, 284 } 285 286 EXPONENT = { 287 TokenType.CARET: exp.Pow, 288 } 289 290 RANGE_PARSERS = { 291 **parser.Parser.RANGE_PARSERS, 292 TokenType.DAMP: binary_range_parser(exp.ArrayOverlaps), 293 TokenType.AT_GT: binary_range_parser(exp.ArrayContains), 294 TokenType.LT_AT: binary_range_parser(exp.ArrayContained), 295 } 296 297 def _parse_factor(self) -> t.Optional[exp.Expression]: 298 return self._parse_tokens(self._parse_exponent, self.FACTOR) 299 300 def _parse_exponent(self) -> t.Optional[exp.Expression]: 301 return self._parse_tokens(self._parse_unary, self.EXPONENT) 302 303 def _parse_date_part(self) -> exp.Expression: 304 part = self._parse_type() 305 self._match(TokenType.COMMA) 306 value = self._parse_bitwise() 307 308 if part and part.is_string: 309 part = exp.var(part.name) 310 311 return self.expression(exp.Extract, this=part, expression=value)
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
Inherited Members
313 class Generator(generator.Generator): 314 SINGLE_STRING_INTERVAL = True 315 LOCKING_READS_SUPPORTED = True 316 JOIN_HINTS = False 317 TABLE_HINTS = False 318 PARAMETER_TOKEN = "$" 319 320 TYPE_MAPPING = { 321 **generator.Generator.TYPE_MAPPING, 322 exp.DataType.Type.TINYINT: "SMALLINT", 323 exp.DataType.Type.FLOAT: "REAL", 324 exp.DataType.Type.DOUBLE: "DOUBLE PRECISION", 325 exp.DataType.Type.BINARY: "BYTEA", 326 exp.DataType.Type.VARBINARY: "BYTEA", 327 exp.DataType.Type.DATETIME: "TIMESTAMP", 328 } 329 330 TRANSFORMS = { 331 **generator.Generator.TRANSFORMS, 332 exp.BitwiseXor: lambda self, e: self.binary(e, "#"), 333 exp.ColumnDef: transforms.preprocess([_auto_increment_to_serial, _serial_to_generated]), 334 exp.Explode: rename_func("UNNEST"), 335 exp.JSONExtract: arrow_json_extract_sql, 336 exp.JSONExtractScalar: arrow_json_extract_scalar_sql, 337 exp.JSONBExtract: lambda self, e: self.binary(e, "#>"), 338 exp.JSONBExtractScalar: lambda self, e: self.binary(e, "#>>"), 339 exp.JSONBContains: lambda self, e: self.binary(e, "?"), 340 exp.Pow: lambda self, e: self.binary(e, "^"), 341 exp.CurrentDate: no_paren_current_date_sql, 342 exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP", 343 exp.DateAdd: _date_add_sql("+"), 344 exp.DateStrToDate: datestrtodate_sql, 345 exp.DateSub: _date_add_sql("-"), 346 exp.DateDiff: _date_diff_sql, 347 exp.LogicalOr: rename_func("BOOL_OR"), 348 exp.LogicalAnd: rename_func("BOOL_AND"), 349 exp.Max: max_or_greatest, 350 exp.Min: min_or_least, 351 exp.ArrayOverlaps: lambda self, e: self.binary(e, "&&"), 352 exp.ArrayContains: lambda self, e: self.binary(e, "@>"), 353 exp.ArrayContained: lambda self, e: self.binary(e, "<@"), 354 exp.Merge: transforms.preprocess([transforms.remove_target_from_merge]), 355 exp.Pivot: no_pivot_sql, 356 exp.RegexpLike: lambda self, e: self.binary(e, "~"), 357 exp.RegexpILike: lambda self, e: self.binary(e, "~*"), 358 exp.StrPosition: str_position_sql, 359 exp.StrToTime: lambda self, e: f"TO_TIMESTAMP({self.sql(e, 'this')}, {self.format_time(e)})", 360 exp.Substring: _substring_sql, 361 exp.TimestampTrunc: timestamptrunc_sql, 362 exp.TimeStrToTime: timestrtotime_sql, 363 exp.TimeToStr: lambda self, e: f"TO_CHAR({self.sql(e, 'this')}, {self.format_time(e)})", 364 exp.TableSample: no_tablesample_sql, 365 exp.ToChar: lambda self, e: self.function_fallback_sql(e), 366 exp.Trim: trim_sql, 367 exp.TryCast: no_trycast_sql, 368 exp.TsOrDsToDate: ts_or_ds_to_date_sql("postgres"), 369 exp.UnixToTime: lambda self, e: f"TO_TIMESTAMP({self.sql(e, 'this')})", 370 exp.DataType: _datatype_sql, 371 exp.GroupConcat: _string_agg_sql, 372 exp.Array: lambda self, e: f"{self.normalize_func('ARRAY')}({self.sql(e.expressions[0])})" 373 if isinstance(seq_get(e.expressions, 0), exp.Select) 374 else f"{self.normalize_func('ARRAY')}[{self.expressions(e, flat=True)}]", 375 } 376 377 PROPERTIES_LOCATION = { 378 **generator.Generator.PROPERTIES_LOCATION, 379 exp.TransientProperty: exp.Properties.Location.UNSUPPORTED, 380 exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED, 381 }
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
@classmethod
def
can_identify(text: str, identify: str | bool = 'safe') -> bool:
247 @classmethod 248 def can_identify(cls, text: str, identify: str | bool = "safe") -> bool: 249 """Checks if text can be identified given an identify option. 250 251 Args: 252 text: The text to check. 253 identify: 254 "always" or `True`: Always returns true. 255 "safe": True if the identifier is case-insensitive. 256 257 Returns: 258 Whether or not the given text can be identified. 259 """ 260 if identify is True or identify == "always": 261 return True 262 263 if identify == "safe": 264 return not cls.case_sensitive(text) 265 266 return False
Checks if text can be identified given an identify option.
Arguments:
- text: The text to check.
- identify: "always" or
True: Always returns true. "safe": True if the identifier is case-insensitive.
Returns:
Whether or not the given text can be identified.
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
- createable_sql
- create_sql
- clone_sql
- describe_sql
- prepend_ctes
- with_sql
- cte_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
- escape_str
- 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
- 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
- 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
- safeconcat_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
- oncluster_sql