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