sqlglot.dialects.mysql

  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    date_add_interval_sql,
 10    datestrtodate_sql,
 11    format_time_lambda,
 12    isnull_to_is_null,
 13    json_keyvalue_comma_sql,
 14    locate_to_strposition,
 15    max_or_greatest,
 16    min_or_least,
 17    no_ilike_sql,
 18    no_paren_current_date_sql,
 19    no_pivot_sql,
 20    no_tablesample_sql,
 21    no_trycast_sql,
 22    parse_date_delta_with_interval,
 23    rename_func,
 24    simplify_literal,
 25    strposition_to_locate_sql,
 26)
 27from sqlglot.helper import seq_get
 28from sqlglot.tokens import TokenType
 29
 30
 31def _show_parser(*args: t.Any, **kwargs: t.Any) -> t.Callable[[MySQL.Parser], exp.Show]:
 32    def _parse(self: MySQL.Parser) -> exp.Show:
 33        return self._parse_show_mysql(*args, **kwargs)
 34
 35    return _parse
 36
 37
 38def _date_trunc_sql(self: MySQL.Generator, expression: exp.DateTrunc) -> str:
 39    expr = self.sql(expression, "this")
 40    unit = expression.text("unit")
 41
 42    if unit == "day":
 43        return f"DATE({expr})"
 44
 45    if unit == "week":
 46        concat = f"CONCAT(YEAR({expr}), ' ', WEEK({expr}, 1), ' 1')"
 47        date_format = "%Y %u %w"
 48    elif unit == "month":
 49        concat = f"CONCAT(YEAR({expr}), ' ', MONTH({expr}), ' 1')"
 50        date_format = "%Y %c %e"
 51    elif unit == "quarter":
 52        concat = f"CONCAT(YEAR({expr}), ' ', QUARTER({expr}) * 3 - 2, ' 1')"
 53        date_format = "%Y %c %e"
 54    elif unit == "year":
 55        concat = f"CONCAT(YEAR({expr}), ' 1 1')"
 56        date_format = "%Y %c %e"
 57    else:
 58        self.unsupported(f"Unexpected interval unit: {unit}")
 59        return f"DATE({expr})"
 60
 61    return f"STR_TO_DATE({concat}, '{date_format}')"
 62
 63
 64def _str_to_date(args: t.List) -> exp.StrToDate:
 65    date_format = MySQL.format_time(seq_get(args, 1))
 66    return exp.StrToDate(this=seq_get(args, 0), format=date_format)
 67
 68
 69def _str_to_date_sql(self: MySQL.Generator, expression: exp.StrToDate | exp.StrToTime) -> str:
 70    date_format = self.format_time(expression)
 71    return f"STR_TO_DATE({self.sql(expression.this)}, {date_format})"
 72
 73
 74def _trim_sql(self: MySQL.Generator, expression: exp.Trim) -> str:
 75    target = self.sql(expression, "this")
 76    trim_type = self.sql(expression, "position")
 77    remove_chars = self.sql(expression, "expression")
 78
 79    # Use TRIM/LTRIM/RTRIM syntax if the expression isn't mysql-specific
 80    if not remove_chars:
 81        return self.trim_sql(expression)
 82
 83    trim_type = f"{trim_type} " if trim_type else ""
 84    remove_chars = f"{remove_chars} " if remove_chars else ""
 85    from_part = "FROM " if trim_type or remove_chars else ""
 86    return f"TRIM({trim_type}{remove_chars}{from_part}{target})"
 87
 88
 89def _date_add_sql(kind: str) -> t.Callable[[MySQL.Generator, exp.DateAdd | exp.DateSub], str]:
 90    def func(self: MySQL.Generator, expression: exp.DateAdd | exp.DateSub) -> str:
 91        this = self.sql(expression, "this")
 92        unit = expression.text("unit").upper() or "DAY"
 93        return f"DATE_{kind}({this}, {self.sql(exp.Interval(this=expression.expression.copy(), unit=unit))})"
 94
 95    return func
 96
 97
 98class MySQL(Dialect):
 99    # https://dev.mysql.com/doc/refman/8.0/en/identifiers.html
100    IDENTIFIERS_CAN_START_WITH_DIGIT = True
101
102    TIME_FORMAT = "'%Y-%m-%d %T'"
103    DPIPE_IS_STRING_CONCAT = False
104    SUPPORTS_USER_DEFINED_TYPES = False
105    SUPPORTS_SEMI_ANTI_JOIN = False
106
107    # https://prestodb.io/docs/current/functions/datetime.html#mysql-date-functions
108    TIME_MAPPING = {
109        "%M": "%B",
110        "%c": "%-m",
111        "%e": "%-d",
112        "%h": "%I",
113        "%i": "%M",
114        "%s": "%S",
115        "%u": "%W",
116        "%k": "%-H",
117        "%l": "%-I",
118        "%T": "%H:%M:%S",
119        "%W": "%a",
120    }
121
122    class Tokenizer(tokens.Tokenizer):
123        QUOTES = ["'", '"']
124        COMMENTS = ["--", "#", ("/*", "*/")]
125        IDENTIFIERS = ["`"]
126        STRING_ESCAPES = ["'", '"', "\\"]
127        BIT_STRINGS = [("b'", "'"), ("B'", "'"), ("0b", "")]
128        HEX_STRINGS = [("x'", "'"), ("X'", "'"), ("0x", "")]
129
130        KEYWORDS = {
131            **tokens.Tokenizer.KEYWORDS,
132            "CHARSET": TokenType.CHARACTER_SET,
133            "ENUM": TokenType.ENUM,
134            "FORCE": TokenType.FORCE,
135            "IGNORE": TokenType.IGNORE,
136            "LOCK TABLES": TokenType.COMMAND,
137            "LONGBLOB": TokenType.LONGBLOB,
138            "LONGTEXT": TokenType.LONGTEXT,
139            "MEDIUMBLOB": TokenType.MEDIUMBLOB,
140            "TINYBLOB": TokenType.TINYBLOB,
141            "TINYTEXT": TokenType.TINYTEXT,
142            "MEDIUMTEXT": TokenType.MEDIUMTEXT,
143            "MEDIUMINT": TokenType.MEDIUMINT,
144            "MEMBER OF": TokenType.MEMBER_OF,
145            "SEPARATOR": TokenType.SEPARATOR,
146            "START": TokenType.BEGIN,
147            "SIGNED": TokenType.BIGINT,
148            "SIGNED INTEGER": TokenType.BIGINT,
149            "UNLOCK TABLES": TokenType.COMMAND,
150            "UNSIGNED": TokenType.UBIGINT,
151            "UNSIGNED INTEGER": TokenType.UBIGINT,
152            "YEAR": TokenType.YEAR,
153            "_ARMSCII8": TokenType.INTRODUCER,
154            "_ASCII": TokenType.INTRODUCER,
155            "_BIG5": TokenType.INTRODUCER,
156            "_BINARY": TokenType.INTRODUCER,
157            "_CP1250": TokenType.INTRODUCER,
158            "_CP1251": TokenType.INTRODUCER,
159            "_CP1256": TokenType.INTRODUCER,
160            "_CP1257": TokenType.INTRODUCER,
161            "_CP850": TokenType.INTRODUCER,
162            "_CP852": TokenType.INTRODUCER,
163            "_CP866": TokenType.INTRODUCER,
164            "_CP932": TokenType.INTRODUCER,
165            "_DEC8": TokenType.INTRODUCER,
166            "_EUCJPMS": TokenType.INTRODUCER,
167            "_EUCKR": TokenType.INTRODUCER,
168            "_GB18030": TokenType.INTRODUCER,
169            "_GB2312": TokenType.INTRODUCER,
170            "_GBK": TokenType.INTRODUCER,
171            "_GEOSTD8": TokenType.INTRODUCER,
172            "_GREEK": TokenType.INTRODUCER,
173            "_HEBREW": TokenType.INTRODUCER,
174            "_HP8": TokenType.INTRODUCER,
175            "_KEYBCS2": TokenType.INTRODUCER,
176            "_KOI8R": TokenType.INTRODUCER,
177            "_KOI8U": TokenType.INTRODUCER,
178            "_LATIN1": TokenType.INTRODUCER,
179            "_LATIN2": TokenType.INTRODUCER,
180            "_LATIN5": TokenType.INTRODUCER,
181            "_LATIN7": TokenType.INTRODUCER,
182            "_MACCE": TokenType.INTRODUCER,
183            "_MACROMAN": TokenType.INTRODUCER,
184            "_SJIS": TokenType.INTRODUCER,
185            "_SWE7": TokenType.INTRODUCER,
186            "_TIS620": TokenType.INTRODUCER,
187            "_UCS2": TokenType.INTRODUCER,
188            "_UJIS": TokenType.INTRODUCER,
189            # https://dev.mysql.com/doc/refman/8.0/en/string-literals.html
190            "_UTF8": TokenType.INTRODUCER,
191            "_UTF16": TokenType.INTRODUCER,
192            "_UTF16LE": TokenType.INTRODUCER,
193            "_UTF32": TokenType.INTRODUCER,
194            "_UTF8MB3": TokenType.INTRODUCER,
195            "_UTF8MB4": TokenType.INTRODUCER,
196            "@@": TokenType.SESSION_PARAMETER,
197        }
198
199        COMMANDS = tokens.Tokenizer.COMMANDS - {TokenType.SHOW}
200
201    class Parser(parser.Parser):
202        FUNC_TOKENS = {
203            *parser.Parser.FUNC_TOKENS,
204            TokenType.DATABASE,
205            TokenType.SCHEMA,
206            TokenType.VALUES,
207        }
208
209        CONJUNCTION = {
210            **parser.Parser.CONJUNCTION,
211            TokenType.DAMP: exp.And,
212            TokenType.XOR: exp.Xor,
213            TokenType.DPIPE: exp.Or,
214        }
215
216        # MySQL uses || as a synonym to the logical OR operator
217        # https://dev.mysql.com/doc/refman/8.0/en/logical-operators.html#operator_or
218        BITWISE = parser.Parser.BITWISE.copy()
219        BITWISE.pop(TokenType.DPIPE)
220
221        TABLE_ALIAS_TOKENS = (
222            parser.Parser.TABLE_ALIAS_TOKENS - parser.Parser.TABLE_INDEX_HINT_TOKENS
223        )
224
225        RANGE_PARSERS = {
226            **parser.Parser.RANGE_PARSERS,
227            TokenType.MEMBER_OF: lambda self, this: self.expression(
228                exp.JSONArrayContains,
229                this=this,
230                expression=self._parse_wrapped(self._parse_expression),
231            ),
232        }
233
234        FUNCTIONS = {
235            **parser.Parser.FUNCTIONS,
236            "DATE_ADD": parse_date_delta_with_interval(exp.DateAdd),
237            "DATE_FORMAT": format_time_lambda(exp.TimeToStr, "mysql"),
238            "DATE_SUB": parse_date_delta_with_interval(exp.DateSub),
239            "INSTR": lambda args: exp.StrPosition(substr=seq_get(args, 1), this=seq_get(args, 0)),
240            "ISNULL": isnull_to_is_null,
241            "LOCATE": locate_to_strposition,
242            "MONTHNAME": lambda args: exp.TimeToStr(
243                this=seq_get(args, 0),
244                format=exp.Literal.string("%B"),
245            ),
246            "STR_TO_DATE": _str_to_date,
247        }
248
249        FUNCTION_PARSERS = {
250            **parser.Parser.FUNCTION_PARSERS,
251            "GROUP_CONCAT": lambda self: self.expression(
252                exp.GroupConcat,
253                this=self._parse_lambda(),
254                separator=self._match(TokenType.SEPARATOR) and self._parse_field(),
255            ),
256            # https://dev.mysql.com/doc/refman/5.7/en/miscellaneous-functions.html#function_values
257            "VALUES": lambda self: self.expression(
258                exp.Anonymous, this="VALUES", expressions=[self._parse_id_var()]
259            ),
260        }
261
262        STATEMENT_PARSERS = {
263            **parser.Parser.STATEMENT_PARSERS,
264            TokenType.SHOW: lambda self: self._parse_show(),
265        }
266
267        SHOW_PARSERS = {
268            "BINARY LOGS": _show_parser("BINARY LOGS"),
269            "MASTER LOGS": _show_parser("BINARY LOGS"),
270            "BINLOG EVENTS": _show_parser("BINLOG EVENTS"),
271            "CHARACTER SET": _show_parser("CHARACTER SET"),
272            "CHARSET": _show_parser("CHARACTER SET"),
273            "COLLATION": _show_parser("COLLATION"),
274            "FULL COLUMNS": _show_parser("COLUMNS", target="FROM", full=True),
275            "COLUMNS": _show_parser("COLUMNS", target="FROM"),
276            "CREATE DATABASE": _show_parser("CREATE DATABASE", target=True),
277            "CREATE EVENT": _show_parser("CREATE EVENT", target=True),
278            "CREATE FUNCTION": _show_parser("CREATE FUNCTION", target=True),
279            "CREATE PROCEDURE": _show_parser("CREATE PROCEDURE", target=True),
280            "CREATE TABLE": _show_parser("CREATE TABLE", target=True),
281            "CREATE TRIGGER": _show_parser("CREATE TRIGGER", target=True),
282            "CREATE VIEW": _show_parser("CREATE VIEW", target=True),
283            "DATABASES": _show_parser("DATABASES"),
284            "ENGINE": _show_parser("ENGINE", target=True),
285            "STORAGE ENGINES": _show_parser("ENGINES"),
286            "ENGINES": _show_parser("ENGINES"),
287            "ERRORS": _show_parser("ERRORS"),
288            "EVENTS": _show_parser("EVENTS"),
289            "FUNCTION CODE": _show_parser("FUNCTION CODE", target=True),
290            "FUNCTION STATUS": _show_parser("FUNCTION STATUS"),
291            "GRANTS": _show_parser("GRANTS", target="FOR"),
292            "INDEX": _show_parser("INDEX", target="FROM"),
293            "MASTER STATUS": _show_parser("MASTER STATUS"),
294            "OPEN TABLES": _show_parser("OPEN TABLES"),
295            "PLUGINS": _show_parser("PLUGINS"),
296            "PROCEDURE CODE": _show_parser("PROCEDURE CODE", target=True),
297            "PROCEDURE STATUS": _show_parser("PROCEDURE STATUS"),
298            "PRIVILEGES": _show_parser("PRIVILEGES"),
299            "FULL PROCESSLIST": _show_parser("PROCESSLIST", full=True),
300            "PROCESSLIST": _show_parser("PROCESSLIST"),
301            "PROFILE": _show_parser("PROFILE"),
302            "PROFILES": _show_parser("PROFILES"),
303            "RELAYLOG EVENTS": _show_parser("RELAYLOG EVENTS"),
304            "REPLICAS": _show_parser("REPLICAS"),
305            "SLAVE HOSTS": _show_parser("REPLICAS"),
306            "REPLICA STATUS": _show_parser("REPLICA STATUS"),
307            "SLAVE STATUS": _show_parser("REPLICA STATUS"),
308            "GLOBAL STATUS": _show_parser("STATUS", global_=True),
309            "SESSION STATUS": _show_parser("STATUS"),
310            "STATUS": _show_parser("STATUS"),
311            "TABLE STATUS": _show_parser("TABLE STATUS"),
312            "FULL TABLES": _show_parser("TABLES", full=True),
313            "TABLES": _show_parser("TABLES"),
314            "TRIGGERS": _show_parser("TRIGGERS"),
315            "GLOBAL VARIABLES": _show_parser("VARIABLES", global_=True),
316            "SESSION VARIABLES": _show_parser("VARIABLES"),
317            "VARIABLES": _show_parser("VARIABLES"),
318            "WARNINGS": _show_parser("WARNINGS"),
319        }
320
321        SET_PARSERS = {
322            **parser.Parser.SET_PARSERS,
323            "PERSIST": lambda self: self._parse_set_item_assignment("PERSIST"),
324            "PERSIST_ONLY": lambda self: self._parse_set_item_assignment("PERSIST_ONLY"),
325            "CHARACTER SET": lambda self: self._parse_set_item_charset("CHARACTER SET"),
326            "CHARSET": lambda self: self._parse_set_item_charset("CHARACTER SET"),
327            "NAMES": lambda self: self._parse_set_item_names(),
328        }
329
330        CONSTRAINT_PARSERS = {
331            **parser.Parser.CONSTRAINT_PARSERS,
332            "FULLTEXT": lambda self: self._parse_index_constraint(kind="FULLTEXT"),
333            "INDEX": lambda self: self._parse_index_constraint(),
334            "KEY": lambda self: self._parse_index_constraint(),
335            "SPATIAL": lambda self: self._parse_index_constraint(kind="SPATIAL"),
336        }
337
338        SCHEMA_UNNAMED_CONSTRAINTS = {
339            *parser.Parser.SCHEMA_UNNAMED_CONSTRAINTS,
340            "FULLTEXT",
341            "INDEX",
342            "KEY",
343            "SPATIAL",
344        }
345
346        PROFILE_TYPES = {
347            "ALL",
348            "BLOCK IO",
349            "CONTEXT SWITCHES",
350            "CPU",
351            "IPC",
352            "MEMORY",
353            "PAGE FAULTS",
354            "SOURCE",
355            "SWAPS",
356        }
357
358        TYPE_TOKENS = {
359            *parser.Parser.TYPE_TOKENS,
360            TokenType.SET,
361        }
362
363        ENUM_TYPE_TOKENS = {
364            *parser.Parser.ENUM_TYPE_TOKENS,
365            TokenType.SET,
366        }
367
368        LOG_DEFAULTS_TO_LN = True
369
370        def _parse_primary_key_part(self) -> t.Optional[exp.Expression]:
371            this = self._parse_id_var()
372            if not self._match(TokenType.L_PAREN):
373                return this
374
375            expression = self._parse_number()
376            self._match_r_paren()
377            return self.expression(exp.ColumnPrefix, this=this, expression=expression)
378
379        def _parse_index_constraint(
380            self, kind: t.Optional[str] = None
381        ) -> exp.IndexColumnConstraint:
382            if kind:
383                self._match_texts({"INDEX", "KEY"})
384
385            this = self._parse_id_var(any_token=False)
386            index_type = self._match(TokenType.USING) and self._advance_any() and self._prev.text
387            schema = self._parse_schema()
388
389            options = []
390            while True:
391                if self._match_text_seq("KEY_BLOCK_SIZE"):
392                    self._match(TokenType.EQ)
393                    opt = exp.IndexConstraintOption(key_block_size=self._parse_number())
394                elif self._match(TokenType.USING):
395                    opt = exp.IndexConstraintOption(using=self._advance_any() and self._prev.text)
396                elif self._match_text_seq("WITH", "PARSER"):
397                    opt = exp.IndexConstraintOption(parser=self._parse_var(any_token=True))
398                elif self._match(TokenType.COMMENT):
399                    opt = exp.IndexConstraintOption(comment=self._parse_string())
400                elif self._match_text_seq("VISIBLE"):
401                    opt = exp.IndexConstraintOption(visible=True)
402                elif self._match_text_seq("INVISIBLE"):
403                    opt = exp.IndexConstraintOption(visible=False)
404                elif self._match_text_seq("ENGINE_ATTRIBUTE"):
405                    self._match(TokenType.EQ)
406                    opt = exp.IndexConstraintOption(engine_attr=self._parse_string())
407                elif self._match_text_seq("ENGINE_ATTRIBUTE"):
408                    self._match(TokenType.EQ)
409                    opt = exp.IndexConstraintOption(engine_attr=self._parse_string())
410                elif self._match_text_seq("SECONDARY_ENGINE_ATTRIBUTE"):
411                    self._match(TokenType.EQ)
412                    opt = exp.IndexConstraintOption(secondary_engine_attr=self._parse_string())
413                else:
414                    opt = None
415
416                if not opt:
417                    break
418
419                options.append(opt)
420
421            return self.expression(
422                exp.IndexColumnConstraint,
423                this=this,
424                schema=schema,
425                kind=kind,
426                index_type=index_type,
427                options=options,
428            )
429
430        def _parse_show_mysql(
431            self,
432            this: str,
433            target: bool | str = False,
434            full: t.Optional[bool] = None,
435            global_: t.Optional[bool] = None,
436        ) -> exp.Show:
437            if target:
438                if isinstance(target, str):
439                    self._match_text_seq(target)
440                target_id = self._parse_id_var()
441            else:
442                target_id = None
443
444            log = self._parse_string() if self._match_text_seq("IN") else None
445
446            if this in {"BINLOG EVENTS", "RELAYLOG EVENTS"}:
447                position = self._parse_number() if self._match_text_seq("FROM") else None
448                db = None
449            else:
450                position = None
451                db = None
452
453                if self._match(TokenType.FROM):
454                    db = self._parse_id_var()
455                elif self._match(TokenType.DOT):
456                    db = target_id
457                    target_id = self._parse_id_var()
458
459            channel = self._parse_id_var() if self._match_text_seq("FOR", "CHANNEL") else None
460
461            like = self._parse_string() if self._match_text_seq("LIKE") else None
462            where = self._parse_where()
463
464            if this == "PROFILE":
465                types = self._parse_csv(lambda: self._parse_var_from_options(self.PROFILE_TYPES))
466                query = self._parse_number() if self._match_text_seq("FOR", "QUERY") else None
467                offset = self._parse_number() if self._match_text_seq("OFFSET") else None
468                limit = self._parse_number() if self._match_text_seq("LIMIT") else None
469            else:
470                types, query = None, None
471                offset, limit = self._parse_oldstyle_limit()
472
473            mutex = True if self._match_text_seq("MUTEX") else None
474            mutex = False if self._match_text_seq("STATUS") else mutex
475
476            return self.expression(
477                exp.Show,
478                this=this,
479                target=target_id,
480                full=full,
481                log=log,
482                position=position,
483                db=db,
484                channel=channel,
485                like=like,
486                where=where,
487                types=types,
488                query=query,
489                offset=offset,
490                limit=limit,
491                mutex=mutex,
492                **{"global": global_},  # type: ignore
493            )
494
495        def _parse_oldstyle_limit(
496            self,
497        ) -> t.Tuple[t.Optional[exp.Expression], t.Optional[exp.Expression]]:
498            limit = None
499            offset = None
500            if self._match_text_seq("LIMIT"):
501                parts = self._parse_csv(self._parse_number)
502                if len(parts) == 1:
503                    limit = parts[0]
504                elif len(parts) == 2:
505                    limit = parts[1]
506                    offset = parts[0]
507
508            return offset, limit
509
510        def _parse_set_item_charset(self, kind: str) -> exp.Expression:
511            this = self._parse_string() or self._parse_id_var()
512            return self.expression(exp.SetItem, this=this, kind=kind)
513
514        def _parse_set_item_names(self) -> exp.Expression:
515            charset = self._parse_string() or self._parse_id_var()
516            if self._match_text_seq("COLLATE"):
517                collate = self._parse_string() or self._parse_id_var()
518            else:
519                collate = None
520
521            return self.expression(exp.SetItem, this=charset, collate=collate, kind="NAMES")
522
523        def _parse_type(self) -> t.Optional[exp.Expression]:
524            # mysql binary is special and can work anywhere, even in order by operations
525            # it operates like a no paren func
526            if self._match(TokenType.BINARY, advance=False):
527                data_type = self._parse_types(check_func=True, allow_identifiers=False)
528
529                if isinstance(data_type, exp.DataType):
530                    return self.expression(exp.Cast, this=self._parse_column(), to=data_type)
531
532            return super()._parse_type()
533
534    class Generator(generator.Generator):
535        LOCKING_READS_SUPPORTED = True
536        NULL_ORDERING_SUPPORTED = False
537        JOIN_HINTS = False
538        TABLE_HINTS = True
539        DUPLICATE_KEY_UPDATE_WITH_SET = False
540        QUERY_HINT_SEP = " "
541        VALUES_AS_TABLE = False
542        NVL2_SUPPORTED = False
543
544        TRANSFORMS = {
545            **generator.Generator.TRANSFORMS,
546            exp.CurrentDate: no_paren_current_date_sql,
547            exp.DateDiff: lambda self, e: self.func("DATEDIFF", e.this, e.expression),
548            exp.DateAdd: _date_add_sql("ADD"),
549            exp.DateStrToDate: datestrtodate_sql,
550            exp.DateSub: _date_add_sql("SUB"),
551            exp.DateTrunc: _date_trunc_sql,
552            exp.DayOfMonth: rename_func("DAYOFMONTH"),
553            exp.DayOfWeek: rename_func("DAYOFWEEK"),
554            exp.DayOfYear: rename_func("DAYOFYEAR"),
555            exp.GroupConcat: lambda self, e: f"""GROUP_CONCAT({self.sql(e, "this")} SEPARATOR {self.sql(e, "separator") or "','"})""",
556            exp.ILike: no_ilike_sql,
557            exp.JSONExtractScalar: arrow_json_extract_scalar_sql,
558            exp.JSONKeyValue: json_keyvalue_comma_sql,
559            exp.Max: max_or_greatest,
560            exp.Min: min_or_least,
561            exp.NullSafeEQ: lambda self, e: self.binary(e, "<=>"),
562            exp.NullSafeNEQ: lambda self, e: self.not_sql(self.binary(e, "<=>")),
563            exp.Pivot: no_pivot_sql,
564            exp.Select: transforms.preprocess(
565                [transforms.eliminate_distinct_on, transforms.eliminate_semi_and_anti_joins]
566            ),
567            exp.StrPosition: strposition_to_locate_sql,
568            exp.StrToDate: _str_to_date_sql,
569            exp.StrToTime: _str_to_date_sql,
570            exp.Stuff: rename_func("INSERT"),
571            exp.TableSample: no_tablesample_sql,
572            exp.TimestampAdd: date_add_interval_sql("DATE", "ADD"),
573            exp.TimestampSub: date_add_interval_sql("DATE", "SUB"),
574            exp.TimeStrToUnix: rename_func("UNIX_TIMESTAMP"),
575            exp.TimeStrToTime: lambda self, e: self.sql(exp.cast(e.this, "datetime", copy=True)),
576            exp.TimeToStr: lambda self, e: self.func("DATE_FORMAT", e.this, self.format_time(e)),
577            exp.Trim: _trim_sql,
578            exp.TryCast: no_trycast_sql,
579            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
580        }
581
582        UNSIGNED_TYPE_MAPPING = {
583            exp.DataType.Type.UBIGINT: "BIGINT",
584            exp.DataType.Type.UINT: "INT",
585            exp.DataType.Type.UMEDIUMINT: "MEDIUMINT",
586            exp.DataType.Type.USMALLINT: "SMALLINT",
587            exp.DataType.Type.UTINYINT: "TINYINT",
588        }
589
590        TIMESTAMP_TYPE_MAPPING = {
591            exp.DataType.Type.TIMESTAMP: "DATETIME",
592            exp.DataType.Type.TIMESTAMPTZ: "TIMESTAMP",
593            exp.DataType.Type.TIMESTAMPLTZ: "TIMESTAMP",
594        }
595
596        TYPE_MAPPING = {
597            **generator.Generator.TYPE_MAPPING,
598            **UNSIGNED_TYPE_MAPPING,
599            **TIMESTAMP_TYPE_MAPPING,
600        }
601
602        TYPE_MAPPING.pop(exp.DataType.Type.MEDIUMTEXT)
603        TYPE_MAPPING.pop(exp.DataType.Type.LONGTEXT)
604        TYPE_MAPPING.pop(exp.DataType.Type.TINYTEXT)
605        TYPE_MAPPING.pop(exp.DataType.Type.MEDIUMBLOB)
606        TYPE_MAPPING.pop(exp.DataType.Type.LONGBLOB)
607        TYPE_MAPPING.pop(exp.DataType.Type.TINYBLOB)
608
609        PROPERTIES_LOCATION = {
610            **generator.Generator.PROPERTIES_LOCATION,
611            exp.TransientProperty: exp.Properties.Location.UNSUPPORTED,
612            exp.VolatileProperty: exp.Properties.Location.UNSUPPORTED,
613        }
614
615        LIMIT_FETCH = "LIMIT"
616
617        # MySQL doesn't support many datatypes in cast.
618        # https://dev.mysql.com/doc/refman/8.0/en/cast-functions.html#function_cast
619        CAST_MAPPING = {
620            exp.DataType.Type.BIGINT: "SIGNED",
621            exp.DataType.Type.BOOLEAN: "SIGNED",
622            exp.DataType.Type.INT: "SIGNED",
623            exp.DataType.Type.TEXT: "CHAR",
624            exp.DataType.Type.UBIGINT: "UNSIGNED",
625            exp.DataType.Type.VARCHAR: "CHAR",
626        }
627
628        TIMESTAMP_FUNC_TYPES = {
629            exp.DataType.Type.TIMESTAMPTZ,
630            exp.DataType.Type.TIMESTAMPLTZ,
631        }
632
633        def datatype_sql(self, expression: exp.DataType) -> str:
634            # https://dev.mysql.com/doc/refman/8.0/en/numeric-type-syntax.html
635            result = super().datatype_sql(expression)
636            if expression.this in self.UNSIGNED_TYPE_MAPPING:
637                result = f"{result} UNSIGNED"
638            return result
639
640        def limit_sql(self, expression: exp.Limit, top: bool = False) -> str:
641            # MySQL requires simple literal values for its LIMIT clause.
642            expression = simplify_literal(expression.copy())
643            return super().limit_sql(expression, top=top)
644
645        def offset_sql(self, expression: exp.Offset) -> str:
646            # MySQL requires simple literal values for its OFFSET clause.
647            expression = simplify_literal(expression.copy())
648            return super().offset_sql(expression)
649
650        def xor_sql(self, expression: exp.Xor) -> str:
651            if expression.expressions:
652                return self.expressions(expression, sep=" XOR ")
653            return super().xor_sql(expression)
654
655        def jsonarraycontains_sql(self, expression: exp.JSONArrayContains) -> str:
656            return f"{self.sql(expression, 'this')} MEMBER OF({self.sql(expression, 'expression')})"
657
658        def cast_sql(self, expression: exp.Cast, safe_prefix: t.Optional[str] = None) -> str:
659            if expression.to.this in self.TIMESTAMP_FUNC_TYPES:
660                return self.func("TIMESTAMP", expression.this)
661
662            to = self.CAST_MAPPING.get(expression.to.this)
663
664            if to:
665                expression = expression.copy()
666                expression.to.set("this", to)
667            return super().cast_sql(expression)
668
669        def show_sql(self, expression: exp.Show) -> str:
670            this = f" {expression.name}"
671            full = " FULL" if expression.args.get("full") else ""
672            global_ = " GLOBAL" if expression.args.get("global") else ""
673
674            target = self.sql(expression, "target")
675            target = f" {target}" if target else ""
676            if expression.name in {"COLUMNS", "INDEX"}:
677                target = f" FROM{target}"
678            elif expression.name == "GRANTS":
679                target = f" FOR{target}"
680
681            db = self._prefixed_sql("FROM", expression, "db")
682
683            like = self._prefixed_sql("LIKE", expression, "like")
684            where = self.sql(expression, "where")
685
686            types = self.expressions(expression, key="types")
687            types = f" {types}" if types else types
688            query = self._prefixed_sql("FOR QUERY", expression, "query")
689
690            if expression.name == "PROFILE":
691                offset = self._prefixed_sql("OFFSET", expression, "offset")
692                limit = self._prefixed_sql("LIMIT", expression, "limit")
693            else:
694                offset = ""
695                limit = self._oldstyle_limit_sql(expression)
696
697            log = self._prefixed_sql("IN", expression, "log")
698            position = self._prefixed_sql("FROM", expression, "position")
699
700            channel = self._prefixed_sql("FOR CHANNEL", expression, "channel")
701
702            if expression.name == "ENGINE":
703                mutex_or_status = " MUTEX" if expression.args.get("mutex") else " STATUS"
704            else:
705                mutex_or_status = ""
706
707            return f"SHOW{full}{global_}{this}{target}{types}{db}{query}{log}{position}{channel}{mutex_or_status}{like}{where}{offset}{limit}"
708
709        def _prefixed_sql(self, prefix: str, expression: exp.Expression, arg: str) -> str:
710            sql = self.sql(expression, arg)
711            return f" {prefix} {sql}" if sql else ""
712
713        def _oldstyle_limit_sql(self, expression: exp.Show) -> str:
714            limit = self.sql(expression, "limit")
715            offset = self.sql(expression, "offset")
716            if limit:
717                limit_offset = f"{offset}, {limit}" if offset else limit
718                return f" LIMIT {limit_offset}"
719            return ""