sqlglot.tokens

   1from __future__ import annotations
   2
   3import os
   4import typing as t
   5from enum import auto
   6
   7from sqlglot.errors import SqlglotError, TokenError
   8from sqlglot.helper import AutoName
   9from sqlglot.trie import TrieResult, in_trie, new_trie
  10
  11if t.TYPE_CHECKING:
  12    from sqlglot.dialects.dialect import DialectType
  13
  14
  15try:
  16    from sqlglotrs import (  # type: ignore
  17        Tokenizer as RsTokenizer,
  18        TokenizerDialectSettings as RsTokenizerDialectSettings,
  19        TokenizerSettings as RsTokenizerSettings,
  20        TokenTypeSettings as RsTokenTypeSettings,
  21    )
  22
  23    USE_RS_TOKENIZER = os.environ.get("SQLGLOTRS_TOKENIZER", "1") == "1"
  24except ImportError:
  25    USE_RS_TOKENIZER = False
  26
  27
  28class TokenType(AutoName):
  29    L_PAREN = auto()
  30    R_PAREN = auto()
  31    L_BRACKET = auto()
  32    R_BRACKET = auto()
  33    L_BRACE = auto()
  34    R_BRACE = auto()
  35    COMMA = auto()
  36    DOT = auto()
  37    DASH = auto()
  38    PLUS = auto()
  39    COLON = auto()
  40    DOTCOLON = auto()
  41    DCOLON = auto()
  42    DQMARK = auto()
  43    SEMICOLON = auto()
  44    STAR = auto()
  45    BACKSLASH = auto()
  46    SLASH = auto()
  47    LT = auto()
  48    LTE = auto()
  49    GT = auto()
  50    GTE = auto()
  51    NOT = auto()
  52    EQ = auto()
  53    NEQ = auto()
  54    NULLSAFE_EQ = auto()
  55    COLON_EQ = auto()
  56    AND = auto()
  57    OR = auto()
  58    AMP = auto()
  59    DPIPE = auto()
  60    PIPE_GT = auto()
  61    PIPE = auto()
  62    PIPE_SLASH = auto()
  63    DPIPE_SLASH = auto()
  64    CARET = auto()
  65    CARET_AT = auto()
  66    TILDA = auto()
  67    ARROW = auto()
  68    DARROW = auto()
  69    FARROW = auto()
  70    HASH = auto()
  71    HASH_ARROW = auto()
  72    DHASH_ARROW = auto()
  73    LR_ARROW = auto()
  74    DAT = auto()
  75    LT_AT = auto()
  76    AT_GT = auto()
  77    DOLLAR = auto()
  78    PARAMETER = auto()
  79    SESSION_PARAMETER = auto()
  80    DAMP = auto()
  81    XOR = auto()
  82    DSTAR = auto()
  83
  84    URI_START = auto()
  85
  86    BLOCK_START = auto()
  87    BLOCK_END = auto()
  88
  89    SPACE = auto()
  90    BREAK = auto()
  91
  92    STRING = auto()
  93    NUMBER = auto()
  94    IDENTIFIER = auto()
  95    DATABASE = auto()
  96    COLUMN = auto()
  97    COLUMN_DEF = auto()
  98    SCHEMA = auto()
  99    TABLE = auto()
 100    WAREHOUSE = auto()
 101    STAGE = auto()
 102    STREAMLIT = auto()
 103    VAR = auto()
 104    BIT_STRING = auto()
 105    HEX_STRING = auto()
 106    BYTE_STRING = auto()
 107    NATIONAL_STRING = auto()
 108    RAW_STRING = auto()
 109    HEREDOC_STRING = auto()
 110    UNICODE_STRING = auto()
 111
 112    # types
 113    BIT = auto()
 114    BOOLEAN = auto()
 115    TINYINT = auto()
 116    UTINYINT = auto()
 117    SMALLINT = auto()
 118    USMALLINT = auto()
 119    MEDIUMINT = auto()
 120    UMEDIUMINT = auto()
 121    INT = auto()
 122    UINT = auto()
 123    BIGINT = auto()
 124    UBIGINT = auto()
 125    INT128 = auto()
 126    UINT128 = auto()
 127    INT256 = auto()
 128    UINT256 = auto()
 129    FLOAT = auto()
 130    DOUBLE = auto()
 131    UDOUBLE = auto()
 132    DECIMAL = auto()
 133    DECIMAL32 = auto()
 134    DECIMAL64 = auto()
 135    DECIMAL128 = auto()
 136    DECIMAL256 = auto()
 137    UDECIMAL = auto()
 138    BIGDECIMAL = auto()
 139    CHAR = auto()
 140    NCHAR = auto()
 141    VARCHAR = auto()
 142    NVARCHAR = auto()
 143    BPCHAR = auto()
 144    TEXT = auto()
 145    MEDIUMTEXT = auto()
 146    LONGTEXT = auto()
 147    BLOB = auto()
 148    MEDIUMBLOB = auto()
 149    LONGBLOB = auto()
 150    TINYBLOB = auto()
 151    TINYTEXT = auto()
 152    NAME = auto()
 153    BINARY = auto()
 154    VARBINARY = auto()
 155    JSON = auto()
 156    JSONB = auto()
 157    TIME = auto()
 158    TIMETZ = auto()
 159    TIMESTAMP = auto()
 160    TIMESTAMPTZ = auto()
 161    TIMESTAMPLTZ = auto()
 162    TIMESTAMPNTZ = auto()
 163    TIMESTAMP_S = auto()
 164    TIMESTAMP_MS = auto()
 165    TIMESTAMP_NS = auto()
 166    DATETIME = auto()
 167    DATETIME2 = auto()
 168    DATETIME64 = auto()
 169    SMALLDATETIME = auto()
 170    DATE = auto()
 171    DATE32 = auto()
 172    INT4RANGE = auto()
 173    INT4MULTIRANGE = auto()
 174    INT8RANGE = auto()
 175    INT8MULTIRANGE = auto()
 176    NUMRANGE = auto()
 177    NUMMULTIRANGE = auto()
 178    TSRANGE = auto()
 179    TSMULTIRANGE = auto()
 180    TSTZRANGE = auto()
 181    TSTZMULTIRANGE = auto()
 182    DATERANGE = auto()
 183    DATEMULTIRANGE = auto()
 184    UUID = auto()
 185    GEOGRAPHY = auto()
 186    NULLABLE = auto()
 187    GEOMETRY = auto()
 188    POINT = auto()
 189    RING = auto()
 190    LINESTRING = auto()
 191    MULTILINESTRING = auto()
 192    POLYGON = auto()
 193    MULTIPOLYGON = auto()
 194    HLLSKETCH = auto()
 195    HSTORE = auto()
 196    SUPER = auto()
 197    SERIAL = auto()
 198    SMALLSERIAL = auto()
 199    BIGSERIAL = auto()
 200    XML = auto()
 201    YEAR = auto()
 202    USERDEFINED = auto()
 203    MONEY = auto()
 204    SMALLMONEY = auto()
 205    ROWVERSION = auto()
 206    IMAGE = auto()
 207    VARIANT = auto()
 208    OBJECT = auto()
 209    INET = auto()
 210    IPADDRESS = auto()
 211    IPPREFIX = auto()
 212    IPV4 = auto()
 213    IPV6 = auto()
 214    ENUM = auto()
 215    ENUM8 = auto()
 216    ENUM16 = auto()
 217    FIXEDSTRING = auto()
 218    LOWCARDINALITY = auto()
 219    NESTED = auto()
 220    AGGREGATEFUNCTION = auto()
 221    SIMPLEAGGREGATEFUNCTION = auto()
 222    TDIGEST = auto()
 223    UNKNOWN = auto()
 224    VECTOR = auto()
 225    DYNAMIC = auto()
 226    VOID = auto()
 227
 228    # keywords
 229    ALIAS = auto()
 230    ALTER = auto()
 231    ALWAYS = auto()
 232    ALL = auto()
 233    ANTI = auto()
 234    ANY = auto()
 235    APPLY = auto()
 236    ARRAY = auto()
 237    ASC = auto()
 238    ASOF = auto()
 239    ATTACH = auto()
 240    AUTO_INCREMENT = auto()
 241    BEGIN = auto()
 242    BETWEEN = auto()
 243    BULK_COLLECT_INTO = auto()
 244    CACHE = auto()
 245    CASE = auto()
 246    CHARACTER_SET = auto()
 247    CLUSTER_BY = auto()
 248    COLLATE = auto()
 249    COMMAND = auto()
 250    COMMENT = auto()
 251    COMMIT = auto()
 252    CONNECT_BY = auto()
 253    CONSTRAINT = auto()
 254    COPY = auto()
 255    CREATE = auto()
 256    CROSS = auto()
 257    CUBE = auto()
 258    CURRENT_DATE = auto()
 259    CURRENT_DATETIME = auto()
 260    CURRENT_SCHEMA = auto()
 261    CURRENT_TIME = auto()
 262    CURRENT_TIMESTAMP = auto()
 263    CURRENT_USER = auto()
 264    DECLARE = auto()
 265    DEFAULT = auto()
 266    DELETE = auto()
 267    DESC = auto()
 268    DESCRIBE = auto()
 269    DETACH = auto()
 270    DICTIONARY = auto()
 271    DISTINCT = auto()
 272    DISTRIBUTE_BY = auto()
 273    DIV = auto()
 274    DROP = auto()
 275    ELSE = auto()
 276    END = auto()
 277    ESCAPE = auto()
 278    EXCEPT = auto()
 279    EXECUTE = auto()
 280    EXISTS = auto()
 281    FALSE = auto()
 282    FETCH = auto()
 283    FILE_FORMAT = auto()
 284    FILTER = auto()
 285    FINAL = auto()
 286    FIRST = auto()
 287    FOR = auto()
 288    FORCE = auto()
 289    FOREIGN_KEY = auto()
 290    FORMAT = auto()
 291    FROM = auto()
 292    FULL = auto()
 293    FUNCTION = auto()
 294    GET = auto()
 295    GLOB = auto()
 296    GLOBAL = auto()
 297    GRANT = auto()
 298    GROUP_BY = auto()
 299    GROUPING_SETS = auto()
 300    HAVING = auto()
 301    HINT = auto()
 302    IGNORE = auto()
 303    ILIKE = auto()
 304    ILIKE_ANY = auto()
 305    IN = auto()
 306    INDEX = auto()
 307    INNER = auto()
 308    INSERT = auto()
 309    INTERSECT = auto()
 310    INTERVAL = auto()
 311    INTO = auto()
 312    INTRODUCER = auto()
 313    IRLIKE = auto()
 314    IS = auto()
 315    ISNULL = auto()
 316    JOIN = auto()
 317    JOIN_MARKER = auto()
 318    KEEP = auto()
 319    KEY = auto()
 320    KILL = auto()
 321    LANGUAGE = auto()
 322    LATERAL = auto()
 323    LEFT = auto()
 324    LIKE = auto()
 325    LIKE_ANY = auto()
 326    LIMIT = auto()
 327    LIST = auto()
 328    LOAD = auto()
 329    LOCK = auto()
 330    MAP = auto()
 331    MATCH_CONDITION = auto()
 332    MATCH_RECOGNIZE = auto()
 333    MEMBER_OF = auto()
 334    MERGE = auto()
 335    MOD = auto()
 336    MODEL = auto()
 337    NATURAL = auto()
 338    NEXT = auto()
 339    NOTHING = auto()
 340    NOTNULL = auto()
 341    NULL = auto()
 342    OBJECT_IDENTIFIER = auto()
 343    OFFSET = auto()
 344    ON = auto()
 345    ONLY = auto()
 346    OPERATOR = auto()
 347    ORDER_BY = auto()
 348    ORDER_SIBLINGS_BY = auto()
 349    ORDERED = auto()
 350    ORDINALITY = auto()
 351    OUTER = auto()
 352    OVER = auto()
 353    OVERLAPS = auto()
 354    OVERWRITE = auto()
 355    PARTITION = auto()
 356    PARTITION_BY = auto()
 357    PERCENT = auto()
 358    PIVOT = auto()
 359    PLACEHOLDER = auto()
 360    POSITIONAL = auto()
 361    PRAGMA = auto()
 362    PREWHERE = auto()
 363    PRIMARY_KEY = auto()
 364    PROCEDURE = auto()
 365    PROPERTIES = auto()
 366    PSEUDO_TYPE = auto()
 367    PUT = auto()
 368    QUALIFY = auto()
 369    QUOTE = auto()
 370    RANGE = auto()
 371    RECURSIVE = auto()
 372    REFRESH = auto()
 373    RENAME = auto()
 374    REPLACE = auto()
 375    RETURNING = auto()
 376    REFERENCES = auto()
 377    RIGHT = auto()
 378    RLIKE = auto()
 379    ROLLBACK = auto()
 380    ROLLUP = auto()
 381    ROW = auto()
 382    ROWS = auto()
 383    SELECT = auto()
 384    SEMI = auto()
 385    SEPARATOR = auto()
 386    SEQUENCE = auto()
 387    SERDE_PROPERTIES = auto()
 388    SET = auto()
 389    SETTINGS = auto()
 390    SHOW = auto()
 391    SIMILAR_TO = auto()
 392    SOME = auto()
 393    SORT_BY = auto()
 394    START_WITH = auto()
 395    STORAGE_INTEGRATION = auto()
 396    STRAIGHT_JOIN = auto()
 397    STRUCT = auto()
 398    SUMMARIZE = auto()
 399    TABLE_SAMPLE = auto()
 400    TAG = auto()
 401    TEMPORARY = auto()
 402    TOP = auto()
 403    THEN = auto()
 404    TRUE = auto()
 405    TRUNCATE = auto()
 406    UNCACHE = auto()
 407    UNION = auto()
 408    UNNEST = auto()
 409    UNPIVOT = auto()
 410    UPDATE = auto()
 411    USE = auto()
 412    USING = auto()
 413    VALUES = auto()
 414    VIEW = auto()
 415    VOLATILE = auto()
 416    WHEN = auto()
 417    WHERE = auto()
 418    WINDOW = auto()
 419    WITH = auto()
 420    UNIQUE = auto()
 421    VERSION_SNAPSHOT = auto()
 422    TIMESTAMP_SNAPSHOT = auto()
 423    OPTION = auto()
 424    SINK = auto()
 425    SOURCE = auto()
 426    ANALYZE = auto()
 427    NAMESPACE = auto()
 428    EXPORT = auto()
 429
 430    # sentinel
 431    HIVE_TOKEN_STREAM = auto()
 432
 433
 434_ALL_TOKEN_TYPES = list(TokenType)
 435_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 436
 437
 438class Token:
 439    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 440
 441    @classmethod
 442    def number(cls, number: int) -> Token:
 443        """Returns a NUMBER token with `number` as its text."""
 444        return cls(TokenType.NUMBER, str(number))
 445
 446    @classmethod
 447    def string(cls, string: str) -> Token:
 448        """Returns a STRING token with `string` as its text."""
 449        return cls(TokenType.STRING, string)
 450
 451    @classmethod
 452    def identifier(cls, identifier: str) -> Token:
 453        """Returns an IDENTIFIER token with `identifier` as its text."""
 454        return cls(TokenType.IDENTIFIER, identifier)
 455
 456    @classmethod
 457    def var(cls, var: str) -> Token:
 458        """Returns an VAR token with `var` as its text."""
 459        return cls(TokenType.VAR, var)
 460
 461    def __init__(
 462        self,
 463        token_type: TokenType,
 464        text: str,
 465        line: int = 1,
 466        col: int = 1,
 467        start: int = 0,
 468        end: int = 0,
 469        comments: t.Optional[t.List[str]] = None,
 470    ) -> None:
 471        """Token initializer.
 472
 473        Args:
 474            token_type: The TokenType Enum.
 475            text: The text of the token.
 476            line: The line that the token ends on.
 477            col: The column that the token ends on.
 478            start: The start index of the token.
 479            end: The ending index of the token.
 480            comments: The comments to attach to the token.
 481        """
 482        self.token_type = token_type
 483        self.text = text
 484        self.line = line
 485        self.col = col
 486        self.start = start
 487        self.end = end
 488        self.comments = [] if comments is None else comments
 489
 490    def __repr__(self) -> str:
 491        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 492        return f"<Token {attributes}>"
 493
 494
 495class _Tokenizer(type):
 496    def __new__(cls, clsname, bases, attrs):
 497        klass = super().__new__(cls, clsname, bases, attrs)
 498
 499        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 500            return dict(
 501                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 502            )
 503
 504        def _quotes_to_format(
 505            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 506        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 507            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 508
 509        klass._QUOTES = _convert_quotes(klass.QUOTES)
 510        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 511
 512        klass._FORMAT_STRINGS = {
 513            **{
 514                p + s: (e, TokenType.NATIONAL_STRING)
 515                for s, e in klass._QUOTES.items()
 516                for p in ("n", "N")
 517            },
 518            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 519            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 520            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 521            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 522            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 523            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 524        }
 525
 526        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 527        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 528        klass._COMMENTS = {
 529            **dict(
 530                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 531                for comment in klass.COMMENTS
 532            ),
 533            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 534        }
 535        if klass.HINT_START in klass.KEYWORDS:
 536            klass._COMMENTS[klass.HINT_START] = "*/"
 537
 538        klass._KEYWORD_TRIE = new_trie(
 539            key.upper()
 540            for key in (
 541                *klass.KEYWORDS,
 542                *klass._COMMENTS,
 543                *klass._QUOTES,
 544                *klass._FORMAT_STRINGS,
 545            )
 546            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 547        )
 548
 549        if USE_RS_TOKENIZER:
 550            settings = RsTokenizerSettings(
 551                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 552                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 553                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 554                numeric_literals=klass.NUMERIC_LITERALS,
 555                identifiers=klass._IDENTIFIERS,
 556                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 557                string_escapes=klass._STRING_ESCAPES,
 558                quotes=klass._QUOTES,
 559                format_strings={
 560                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 561                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 562                },
 563                has_bit_strings=bool(klass.BIT_STRINGS),
 564                has_hex_strings=bool(klass.HEX_STRINGS),
 565                comments=klass._COMMENTS,
 566                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 567                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 568                command_prefix_tokens={
 569                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 570                },
 571                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 572                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 573                nested_comments=klass.NESTED_COMMENTS,
 574                hint_start=klass.HINT_START,
 575                tokens_preceding_hint={
 576                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 577                },
 578            )
 579            token_types = RsTokenTypeSettings(
 580                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 581                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 582                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 583                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 584                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 585                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 586                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 587                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 588                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 589                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 590                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 591                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 592                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 593                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 594            )
 595            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 596        else:
 597            klass._RS_TOKENIZER = None
 598
 599        return klass
 600
 601
 602class Tokenizer(metaclass=_Tokenizer):
 603    SINGLE_TOKENS = {
 604        "(": TokenType.L_PAREN,
 605        ")": TokenType.R_PAREN,
 606        "[": TokenType.L_BRACKET,
 607        "]": TokenType.R_BRACKET,
 608        "{": TokenType.L_BRACE,
 609        "}": TokenType.R_BRACE,
 610        "&": TokenType.AMP,
 611        "^": TokenType.CARET,
 612        ":": TokenType.COLON,
 613        ",": TokenType.COMMA,
 614        ".": TokenType.DOT,
 615        "-": TokenType.DASH,
 616        "=": TokenType.EQ,
 617        ">": TokenType.GT,
 618        "<": TokenType.LT,
 619        "%": TokenType.MOD,
 620        "!": TokenType.NOT,
 621        "|": TokenType.PIPE,
 622        "+": TokenType.PLUS,
 623        ";": TokenType.SEMICOLON,
 624        "/": TokenType.SLASH,
 625        "\\": TokenType.BACKSLASH,
 626        "*": TokenType.STAR,
 627        "~": TokenType.TILDA,
 628        "?": TokenType.PLACEHOLDER,
 629        "@": TokenType.PARAMETER,
 630        "#": TokenType.HASH,
 631        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 632        "'": TokenType.UNKNOWN,
 633        "`": TokenType.UNKNOWN,
 634        '"': TokenType.UNKNOWN,
 635    }
 636
 637    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 638    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 639    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 640    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 641    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 642    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 643    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 644    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 645    STRING_ESCAPES = ["'"]
 646    VAR_SINGLE_TOKENS: t.Set[str] = set()
 647
 648    # The strings in this list can always be used as escapes, regardless of the surrounding
 649    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 650    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 651    IDENTIFIER_ESCAPES: t.List[str] = []
 652
 653    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 654    HEREDOC_TAG_IS_IDENTIFIER = False
 655
 656    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 657    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 658
 659    # Whether string escape characters function as such when placed within raw strings
 660    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 661
 662    NESTED_COMMENTS = True
 663
 664    HINT_START = "/*+"
 665
 666    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 667
 668    # Autofilled
 669    _COMMENTS: t.Dict[str, str] = {}
 670    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 671    _IDENTIFIERS: t.Dict[str, str] = {}
 672    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 673    _QUOTES: t.Dict[str, str] = {}
 674    _STRING_ESCAPES: t.Set[str] = set()
 675    _KEYWORD_TRIE: t.Dict = {}
 676    _RS_TOKENIZER: t.Optional[t.Any] = None
 677
 678    KEYWORDS: t.Dict[str, TokenType] = {
 679        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 680        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 681        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 682        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 683        HINT_START: TokenType.HINT,
 684        "==": TokenType.EQ,
 685        "::": TokenType.DCOLON,
 686        "||": TokenType.DPIPE,
 687        "|>": TokenType.PIPE_GT,
 688        ">=": TokenType.GTE,
 689        "<=": TokenType.LTE,
 690        "<>": TokenType.NEQ,
 691        "!=": TokenType.NEQ,
 692        ":=": TokenType.COLON_EQ,
 693        "<=>": TokenType.NULLSAFE_EQ,
 694        "->": TokenType.ARROW,
 695        "->>": TokenType.DARROW,
 696        "=>": TokenType.FARROW,
 697        "#>": TokenType.HASH_ARROW,
 698        "#>>": TokenType.DHASH_ARROW,
 699        "<->": TokenType.LR_ARROW,
 700        "&&": TokenType.DAMP,
 701        "??": TokenType.DQMARK,
 702        "~~~": TokenType.GLOB,
 703        "~~": TokenType.LIKE,
 704        "~~*": TokenType.ILIKE,
 705        "~*": TokenType.IRLIKE,
 706        "ALL": TokenType.ALL,
 707        "ALWAYS": TokenType.ALWAYS,
 708        "AND": TokenType.AND,
 709        "ANTI": TokenType.ANTI,
 710        "ANY": TokenType.ANY,
 711        "ASC": TokenType.ASC,
 712        "AS": TokenType.ALIAS,
 713        "ASOF": TokenType.ASOF,
 714        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 715        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 716        "BEGIN": TokenType.BEGIN,
 717        "BETWEEN": TokenType.BETWEEN,
 718        "CACHE": TokenType.CACHE,
 719        "UNCACHE": TokenType.UNCACHE,
 720        "CASE": TokenType.CASE,
 721        "CHARACTER SET": TokenType.CHARACTER_SET,
 722        "CLUSTER BY": TokenType.CLUSTER_BY,
 723        "COLLATE": TokenType.COLLATE,
 724        "COLUMN": TokenType.COLUMN,
 725        "COMMIT": TokenType.COMMIT,
 726        "CONNECT BY": TokenType.CONNECT_BY,
 727        "CONSTRAINT": TokenType.CONSTRAINT,
 728        "COPY": TokenType.COPY,
 729        "CREATE": TokenType.CREATE,
 730        "CROSS": TokenType.CROSS,
 731        "CUBE": TokenType.CUBE,
 732        "CURRENT_DATE": TokenType.CURRENT_DATE,
 733        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 734        "CURRENT_TIME": TokenType.CURRENT_TIME,
 735        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 736        "CURRENT_USER": TokenType.CURRENT_USER,
 737        "DATABASE": TokenType.DATABASE,
 738        "DEFAULT": TokenType.DEFAULT,
 739        "DELETE": TokenType.DELETE,
 740        "DESC": TokenType.DESC,
 741        "DESCRIBE": TokenType.DESCRIBE,
 742        "DISTINCT": TokenType.DISTINCT,
 743        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 744        "DIV": TokenType.DIV,
 745        "DROP": TokenType.DROP,
 746        "ELSE": TokenType.ELSE,
 747        "END": TokenType.END,
 748        "ENUM": TokenType.ENUM,
 749        "ESCAPE": TokenType.ESCAPE,
 750        "EXCEPT": TokenType.EXCEPT,
 751        "EXECUTE": TokenType.EXECUTE,
 752        "EXISTS": TokenType.EXISTS,
 753        "FALSE": TokenType.FALSE,
 754        "FETCH": TokenType.FETCH,
 755        "FILTER": TokenType.FILTER,
 756        "FIRST": TokenType.FIRST,
 757        "FULL": TokenType.FULL,
 758        "FUNCTION": TokenType.FUNCTION,
 759        "FOR": TokenType.FOR,
 760        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 761        "FORMAT": TokenType.FORMAT,
 762        "FROM": TokenType.FROM,
 763        "GEOGRAPHY": TokenType.GEOGRAPHY,
 764        "GEOMETRY": TokenType.GEOMETRY,
 765        "GLOB": TokenType.GLOB,
 766        "GROUP BY": TokenType.GROUP_BY,
 767        "GROUPING SETS": TokenType.GROUPING_SETS,
 768        "HAVING": TokenType.HAVING,
 769        "ILIKE": TokenType.ILIKE,
 770        "IN": TokenType.IN,
 771        "INDEX": TokenType.INDEX,
 772        "INET": TokenType.INET,
 773        "INNER": TokenType.INNER,
 774        "INSERT": TokenType.INSERT,
 775        "INTERVAL": TokenType.INTERVAL,
 776        "INTERSECT": TokenType.INTERSECT,
 777        "INTO": TokenType.INTO,
 778        "IS": TokenType.IS,
 779        "ISNULL": TokenType.ISNULL,
 780        "JOIN": TokenType.JOIN,
 781        "KEEP": TokenType.KEEP,
 782        "KILL": TokenType.KILL,
 783        "LATERAL": TokenType.LATERAL,
 784        "LEFT": TokenType.LEFT,
 785        "LIKE": TokenType.LIKE,
 786        "LIMIT": TokenType.LIMIT,
 787        "LOAD": TokenType.LOAD,
 788        "LOCK": TokenType.LOCK,
 789        "MERGE": TokenType.MERGE,
 790        "NAMESPACE": TokenType.NAMESPACE,
 791        "NATURAL": TokenType.NATURAL,
 792        "NEXT": TokenType.NEXT,
 793        "NOT": TokenType.NOT,
 794        "NOTNULL": TokenType.NOTNULL,
 795        "NULL": TokenType.NULL,
 796        "OBJECT": TokenType.OBJECT,
 797        "OFFSET": TokenType.OFFSET,
 798        "ON": TokenType.ON,
 799        "OR": TokenType.OR,
 800        "XOR": TokenType.XOR,
 801        "ORDER BY": TokenType.ORDER_BY,
 802        "ORDINALITY": TokenType.ORDINALITY,
 803        "OUTER": TokenType.OUTER,
 804        "OVER": TokenType.OVER,
 805        "OVERLAPS": TokenType.OVERLAPS,
 806        "OVERWRITE": TokenType.OVERWRITE,
 807        "PARTITION": TokenType.PARTITION,
 808        "PARTITION BY": TokenType.PARTITION_BY,
 809        "PARTITIONED BY": TokenType.PARTITION_BY,
 810        "PARTITIONED_BY": TokenType.PARTITION_BY,
 811        "PERCENT": TokenType.PERCENT,
 812        "PIVOT": TokenType.PIVOT,
 813        "PRAGMA": TokenType.PRAGMA,
 814        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 815        "PROCEDURE": TokenType.PROCEDURE,
 816        "QUALIFY": TokenType.QUALIFY,
 817        "RANGE": TokenType.RANGE,
 818        "RECURSIVE": TokenType.RECURSIVE,
 819        "REGEXP": TokenType.RLIKE,
 820        "RENAME": TokenType.RENAME,
 821        "REPLACE": TokenType.REPLACE,
 822        "RETURNING": TokenType.RETURNING,
 823        "REFERENCES": TokenType.REFERENCES,
 824        "RIGHT": TokenType.RIGHT,
 825        "RLIKE": TokenType.RLIKE,
 826        "ROLLBACK": TokenType.ROLLBACK,
 827        "ROLLUP": TokenType.ROLLUP,
 828        "ROW": TokenType.ROW,
 829        "ROWS": TokenType.ROWS,
 830        "SCHEMA": TokenType.SCHEMA,
 831        "SELECT": TokenType.SELECT,
 832        "SEMI": TokenType.SEMI,
 833        "SET": TokenType.SET,
 834        "SETTINGS": TokenType.SETTINGS,
 835        "SHOW": TokenType.SHOW,
 836        "SIMILAR TO": TokenType.SIMILAR_TO,
 837        "SOME": TokenType.SOME,
 838        "SORT BY": TokenType.SORT_BY,
 839        "START WITH": TokenType.START_WITH,
 840        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 841        "TABLE": TokenType.TABLE,
 842        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 843        "TEMP": TokenType.TEMPORARY,
 844        "TEMPORARY": TokenType.TEMPORARY,
 845        "THEN": TokenType.THEN,
 846        "TRUE": TokenType.TRUE,
 847        "TRUNCATE": TokenType.TRUNCATE,
 848        "UNION": TokenType.UNION,
 849        "UNKNOWN": TokenType.UNKNOWN,
 850        "UNNEST": TokenType.UNNEST,
 851        "UNPIVOT": TokenType.UNPIVOT,
 852        "UPDATE": TokenType.UPDATE,
 853        "USE": TokenType.USE,
 854        "USING": TokenType.USING,
 855        "UUID": TokenType.UUID,
 856        "VALUES": TokenType.VALUES,
 857        "VIEW": TokenType.VIEW,
 858        "VOLATILE": TokenType.VOLATILE,
 859        "WHEN": TokenType.WHEN,
 860        "WHERE": TokenType.WHERE,
 861        "WINDOW": TokenType.WINDOW,
 862        "WITH": TokenType.WITH,
 863        "APPLY": TokenType.APPLY,
 864        "ARRAY": TokenType.ARRAY,
 865        "BIT": TokenType.BIT,
 866        "BOOL": TokenType.BOOLEAN,
 867        "BOOLEAN": TokenType.BOOLEAN,
 868        "BYTE": TokenType.TINYINT,
 869        "MEDIUMINT": TokenType.MEDIUMINT,
 870        "INT1": TokenType.TINYINT,
 871        "TINYINT": TokenType.TINYINT,
 872        "INT16": TokenType.SMALLINT,
 873        "SHORT": TokenType.SMALLINT,
 874        "SMALLINT": TokenType.SMALLINT,
 875        "HUGEINT": TokenType.INT128,
 876        "UHUGEINT": TokenType.UINT128,
 877        "INT2": TokenType.SMALLINT,
 878        "INTEGER": TokenType.INT,
 879        "INT": TokenType.INT,
 880        "INT4": TokenType.INT,
 881        "INT32": TokenType.INT,
 882        "INT64": TokenType.BIGINT,
 883        "INT128": TokenType.INT128,
 884        "INT256": TokenType.INT256,
 885        "LONG": TokenType.BIGINT,
 886        "BIGINT": TokenType.BIGINT,
 887        "INT8": TokenType.TINYINT,
 888        "UINT": TokenType.UINT,
 889        "UINT128": TokenType.UINT128,
 890        "UINT256": TokenType.UINT256,
 891        "DEC": TokenType.DECIMAL,
 892        "DECIMAL": TokenType.DECIMAL,
 893        "DECIMAL32": TokenType.DECIMAL32,
 894        "DECIMAL64": TokenType.DECIMAL64,
 895        "DECIMAL128": TokenType.DECIMAL128,
 896        "DECIMAL256": TokenType.DECIMAL256,
 897        "BIGDECIMAL": TokenType.BIGDECIMAL,
 898        "BIGNUMERIC": TokenType.BIGDECIMAL,
 899        "LIST": TokenType.LIST,
 900        "MAP": TokenType.MAP,
 901        "NULLABLE": TokenType.NULLABLE,
 902        "NUMBER": TokenType.DECIMAL,
 903        "NUMERIC": TokenType.DECIMAL,
 904        "FIXED": TokenType.DECIMAL,
 905        "REAL": TokenType.FLOAT,
 906        "FLOAT": TokenType.FLOAT,
 907        "FLOAT4": TokenType.FLOAT,
 908        "FLOAT8": TokenType.DOUBLE,
 909        "DOUBLE": TokenType.DOUBLE,
 910        "DOUBLE PRECISION": TokenType.DOUBLE,
 911        "JSON": TokenType.JSON,
 912        "JSONB": TokenType.JSONB,
 913        "CHAR": TokenType.CHAR,
 914        "CHARACTER": TokenType.CHAR,
 915        "CHAR VARYING": TokenType.VARCHAR,
 916        "CHARACTER VARYING": TokenType.VARCHAR,
 917        "NCHAR": TokenType.NCHAR,
 918        "VARCHAR": TokenType.VARCHAR,
 919        "VARCHAR2": TokenType.VARCHAR,
 920        "NVARCHAR": TokenType.NVARCHAR,
 921        "NVARCHAR2": TokenType.NVARCHAR,
 922        "BPCHAR": TokenType.BPCHAR,
 923        "STR": TokenType.TEXT,
 924        "STRING": TokenType.TEXT,
 925        "TEXT": TokenType.TEXT,
 926        "LONGTEXT": TokenType.LONGTEXT,
 927        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 928        "TINYTEXT": TokenType.TINYTEXT,
 929        "CLOB": TokenType.TEXT,
 930        "LONGVARCHAR": TokenType.TEXT,
 931        "BINARY": TokenType.BINARY,
 932        "BLOB": TokenType.VARBINARY,
 933        "LONGBLOB": TokenType.LONGBLOB,
 934        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 935        "TINYBLOB": TokenType.TINYBLOB,
 936        "BYTEA": TokenType.VARBINARY,
 937        "VARBINARY": TokenType.VARBINARY,
 938        "TIME": TokenType.TIME,
 939        "TIMETZ": TokenType.TIMETZ,
 940        "TIMESTAMP": TokenType.TIMESTAMP,
 941        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 942        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 943        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 944        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 945        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 946        "DATE": TokenType.DATE,
 947        "DATETIME": TokenType.DATETIME,
 948        "INT4RANGE": TokenType.INT4RANGE,
 949        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 950        "INT8RANGE": TokenType.INT8RANGE,
 951        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 952        "NUMRANGE": TokenType.NUMRANGE,
 953        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 954        "TSRANGE": TokenType.TSRANGE,
 955        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 956        "TSTZRANGE": TokenType.TSTZRANGE,
 957        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 958        "DATERANGE": TokenType.DATERANGE,
 959        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 960        "UNIQUE": TokenType.UNIQUE,
 961        "VECTOR": TokenType.VECTOR,
 962        "STRUCT": TokenType.STRUCT,
 963        "SEQUENCE": TokenType.SEQUENCE,
 964        "VARIANT": TokenType.VARIANT,
 965        "ALTER": TokenType.ALTER,
 966        "ANALYZE": TokenType.ANALYZE,
 967        "CALL": TokenType.COMMAND,
 968        "COMMENT": TokenType.COMMENT,
 969        "EXPLAIN": TokenType.COMMAND,
 970        "GRANT": TokenType.GRANT,
 971        "OPTIMIZE": TokenType.COMMAND,
 972        "PREPARE": TokenType.COMMAND,
 973        "VACUUM": TokenType.COMMAND,
 974        "USER-DEFINED": TokenType.USERDEFINED,
 975        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 976        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 977    }
 978
 979    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 980        " ": TokenType.SPACE,
 981        "\t": TokenType.SPACE,
 982        "\n": TokenType.BREAK,
 983        "\r": TokenType.BREAK,
 984    }
 985
 986    COMMANDS = {
 987        TokenType.COMMAND,
 988        TokenType.EXECUTE,
 989        TokenType.FETCH,
 990        TokenType.SHOW,
 991        TokenType.RENAME,
 992    }
 993
 994    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 995
 996    # Handle numeric literals like in hive (3L = BIGINT)
 997    NUMERIC_LITERALS: t.Dict[str, str] = {}
 998
 999    COMMENTS = ["--", ("/*", "*/")]
1000
1001    __slots__ = (
1002        "sql",
1003        "size",
1004        "tokens",
1005        "dialect",
1006        "use_rs_tokenizer",
1007        "_start",
1008        "_current",
1009        "_line",
1010        "_col",
1011        "_comments",
1012        "_char",
1013        "_end",
1014        "_peek",
1015        "_prev_token_line",
1016        "_rs_dialect_settings",
1017    )
1018
1019    def __init__(
1020        self,
1021        dialect: DialectType = None,
1022        use_rs_tokenizer: t.Optional[bool] = None,
1023        **opts: t.Any,
1024    ) -> None:
1025        from sqlglot.dialects import Dialect
1026
1027        self.dialect = Dialect.get_or_raise(dialect)
1028
1029        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1030        self.use_rs_tokenizer = (
1031            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1032        )
1033
1034        if self.use_rs_tokenizer:
1035            self._rs_dialect_settings = RsTokenizerDialectSettings(
1036                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1037                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1038                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1039            )
1040
1041        self.reset()
1042
1043    def reset(self) -> None:
1044        self.sql = ""
1045        self.size = 0
1046        self.tokens: t.List[Token] = []
1047        self._start = 0
1048        self._current = 0
1049        self._line = 1
1050        self._col = 0
1051        self._comments: t.List[str] = []
1052
1053        self._char = ""
1054        self._end = False
1055        self._peek = ""
1056        self._prev_token_line = -1
1057
1058    def tokenize(self, sql: str) -> t.List[Token]:
1059        """Returns a list of tokens corresponding to the SQL string `sql`."""
1060        if self.use_rs_tokenizer:
1061            return self.tokenize_rs(sql)
1062
1063        self.reset()
1064        self.sql = sql
1065        self.size = len(sql)
1066
1067        try:
1068            self._scan()
1069        except Exception as e:
1070            start = max(self._current - 50, 0)
1071            end = min(self._current + 50, self.size - 1)
1072            context = self.sql[start:end]
1073            raise TokenError(f"Error tokenizing '{context}'") from e
1074
1075        return self.tokens
1076
1077    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1078        while self.size and not self._end:
1079            current = self._current
1080
1081            # Skip spaces here rather than iteratively calling advance() for performance reasons
1082            while current < self.size:
1083                char = self.sql[current]
1084
1085                if char.isspace() and (char == " " or char == "\t"):
1086                    current += 1
1087                else:
1088                    break
1089
1090            offset = current - self._current if current > self._current else 1
1091
1092            self._start = current
1093            self._advance(offset)
1094
1095            if not self._char.isspace():
1096                if self._char.isdigit():
1097                    self._scan_number()
1098                elif self._char in self._IDENTIFIERS:
1099                    self._scan_identifier(self._IDENTIFIERS[self._char])
1100                else:
1101                    self._scan_keywords()
1102
1103            if until and until():
1104                break
1105
1106        if self.tokens and self._comments:
1107            self.tokens[-1].comments.extend(self._comments)
1108
1109    def _chars(self, size: int) -> str:
1110        if size == 1:
1111            return self._char
1112
1113        start = self._current - 1
1114        end = start + size
1115
1116        return self.sql[start:end] if end <= self.size else ""
1117
1118    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1119        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1120            # Ensures we don't count an extra line if we get a \r\n line break sequence
1121            if not (self._char == "\r" and self._peek == "\n"):
1122                self._col = i
1123                self._line += 1
1124        else:
1125            self._col += i
1126
1127        self._current += i
1128        self._end = self._current >= self.size
1129        self._char = self.sql[self._current - 1]
1130        self._peek = "" if self._end else self.sql[self._current]
1131
1132        if alnum and self._char.isalnum():
1133            # Here we use local variables instead of attributes for better performance
1134            _col = self._col
1135            _current = self._current
1136            _end = self._end
1137            _peek = self._peek
1138
1139            while _peek.isalnum():
1140                _col += 1
1141                _current += 1
1142                _end = _current >= self.size
1143                _peek = "" if _end else self.sql[_current]
1144
1145            self._col = _col
1146            self._current = _current
1147            self._end = _end
1148            self._peek = _peek
1149            self._char = self.sql[_current - 1]
1150
1151    @property
1152    def _text(self) -> str:
1153        return self.sql[self._start : self._current]
1154
1155    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1156        self._prev_token_line = self._line
1157
1158        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1159            self.tokens[-1].comments.extend(self._comments)
1160            self._comments = []
1161
1162        self.tokens.append(
1163            Token(
1164                token_type,
1165                text=self._text if text is None else text,
1166                line=self._line,
1167                col=self._col,
1168                start=self._start,
1169                end=self._current - 1,
1170                comments=self._comments,
1171            )
1172        )
1173        self._comments = []
1174
1175        # If we have either a semicolon or a begin token before the command's token, we'll parse
1176        # whatever follows the command's token as a string
1177        if (
1178            token_type in self.COMMANDS
1179            and self._peek != ";"
1180            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1181        ):
1182            start = self._current
1183            tokens = len(self.tokens)
1184            self._scan(lambda: self._peek == ";")
1185            self.tokens = self.tokens[:tokens]
1186            text = self.sql[start : self._current].strip()
1187            if text:
1188                self._add(TokenType.STRING, text)
1189
1190    def _scan_keywords(self) -> None:
1191        size = 0
1192        word = None
1193        chars = self._text
1194        char = chars
1195        prev_space = False
1196        skip = False
1197        trie = self._KEYWORD_TRIE
1198        single_token = char in self.SINGLE_TOKENS
1199
1200        while chars:
1201            if skip:
1202                result = TrieResult.PREFIX
1203            else:
1204                result, trie = in_trie(trie, char.upper())
1205
1206            if result == TrieResult.FAILED:
1207                break
1208            if result == TrieResult.EXISTS:
1209                word = chars
1210
1211            end = self._current + size
1212            size += 1
1213
1214            if end < self.size:
1215                char = self.sql[end]
1216                single_token = single_token or char in self.SINGLE_TOKENS
1217                is_space = char.isspace()
1218
1219                if not is_space or not prev_space:
1220                    if is_space:
1221                        char = " "
1222                    chars += char
1223                    prev_space = is_space
1224                    skip = False
1225                else:
1226                    skip = True
1227            else:
1228                char = ""
1229                break
1230
1231        if word:
1232            if self._scan_string(word):
1233                return
1234            if self._scan_comment(word):
1235                return
1236            if prev_space or single_token or not char:
1237                self._advance(size - 1)
1238                word = word.upper()
1239                self._add(self.KEYWORDS[word], text=word)
1240                return
1241
1242        if self._char in self.SINGLE_TOKENS:
1243            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1244            return
1245
1246        self._scan_var()
1247
1248    def _scan_comment(self, comment_start: str) -> bool:
1249        if comment_start not in self._COMMENTS:
1250            return False
1251
1252        comment_start_line = self._line
1253        comment_start_size = len(comment_start)
1254        comment_end = self._COMMENTS[comment_start]
1255
1256        if comment_end:
1257            # Skip the comment's start delimiter
1258            self._advance(comment_start_size)
1259
1260            comment_count = 1
1261            comment_end_size = len(comment_end)
1262
1263            while not self._end:
1264                if self._chars(comment_end_size) == comment_end:
1265                    comment_count -= 1
1266                    if not comment_count:
1267                        break
1268
1269                self._advance(alnum=True)
1270
1271                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1272                if (
1273                    self.NESTED_COMMENTS
1274                    and not self._end
1275                    and self._chars(comment_end_size) == comment_start
1276                ):
1277                    self._advance(comment_start_size)
1278                    comment_count += 1
1279
1280            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1281            self._advance(comment_end_size - 1)
1282        else:
1283            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1284                self._advance(alnum=True)
1285            self._comments.append(self._text[comment_start_size:])
1286
1287        if (
1288            comment_start == self.HINT_START
1289            and self.tokens
1290            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1291        ):
1292            self._add(TokenType.HINT)
1293
1294        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1295        # Multiple consecutive comments are preserved by appending them to the current comments list.
1296        if comment_start_line == self._prev_token_line:
1297            self.tokens[-1].comments.extend(self._comments)
1298            self._comments = []
1299            self._prev_token_line = self._line
1300
1301        return True
1302
1303    def _scan_number(self) -> None:
1304        if self._char == "0":
1305            peek = self._peek.upper()
1306            if peek == "B":
1307                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1308            elif peek == "X":
1309                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1310
1311        decimal = False
1312        scientific = 0
1313
1314        while True:
1315            if self._peek.isdigit():
1316                self._advance()
1317            elif self._peek == "." and not decimal:
1318                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1319                    return self._add(TokenType.NUMBER)
1320                decimal = True
1321                self._advance()
1322            elif self._peek in ("-", "+") and scientific == 1:
1323                scientific += 1
1324                self._advance()
1325            elif self._peek.upper() == "E" and not scientific:
1326                scientific += 1
1327                self._advance()
1328            elif self._peek.isidentifier():
1329                number_text = self._text
1330                literal = ""
1331
1332                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1333                    literal += self._peek
1334                    self._advance()
1335
1336                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1337
1338                if token_type:
1339                    self._add(TokenType.NUMBER, number_text)
1340                    self._add(TokenType.DCOLON, "::")
1341                    return self._add(token_type, literal)
1342                else:
1343                    replaced = literal.replace("_", "")
1344                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1345                        return self._add(TokenType.NUMBER, number_text + replaced)
1346                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1347                        return self._add(TokenType.VAR)
1348
1349                self._advance(-len(literal))
1350                return self._add(TokenType.NUMBER, number_text)
1351            else:
1352                return self._add(TokenType.NUMBER)
1353
1354    def _scan_bits(self) -> None:
1355        self._advance()
1356        value = self._extract_value()
1357        try:
1358            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1359            int(value, 2)
1360            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1361        except ValueError:
1362            self._add(TokenType.IDENTIFIER)
1363
1364    def _scan_hex(self) -> None:
1365        self._advance()
1366        value = self._extract_value()
1367        try:
1368            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1369            int(value, 16)
1370            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1371        except ValueError:
1372            self._add(TokenType.IDENTIFIER)
1373
1374    def _extract_value(self) -> str:
1375        while True:
1376            char = self._peek.strip()
1377            if char and char not in self.SINGLE_TOKENS:
1378                self._advance(alnum=True)
1379            else:
1380                break
1381
1382        return self._text
1383
1384    def _scan_string(self, start: str) -> bool:
1385        base = None
1386        token_type = TokenType.STRING
1387
1388        if start in self._QUOTES:
1389            end = self._QUOTES[start]
1390        elif start in self._FORMAT_STRINGS:
1391            end, token_type = self._FORMAT_STRINGS[start]
1392
1393            if token_type == TokenType.HEX_STRING:
1394                base = 16
1395            elif token_type == TokenType.BIT_STRING:
1396                base = 2
1397            elif token_type == TokenType.HEREDOC_STRING:
1398                self._advance()
1399
1400                if self._char == end:
1401                    tag = ""
1402                else:
1403                    tag = self._extract_string(
1404                        end,
1405                        raw_string=True,
1406                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1407                    )
1408
1409                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1410                    if not self._end:
1411                        self._advance(-1)
1412
1413                    self._advance(-len(tag))
1414                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1415                    return True
1416
1417                end = f"{start}{tag}{end}"
1418        else:
1419            return False
1420
1421        self._advance(len(start))
1422        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1423
1424        if base:
1425            try:
1426                int(text, base)
1427            except Exception:
1428                raise TokenError(
1429                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1430                )
1431
1432        self._add(token_type, text)
1433        return True
1434
1435    def _scan_identifier(self, identifier_end: str) -> None:
1436        self._advance()
1437        text = self._extract_string(
1438            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1439        )
1440        self._add(TokenType.IDENTIFIER, text)
1441
1442    def _scan_var(self) -> None:
1443        while True:
1444            char = self._peek.strip()
1445            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1446                self._advance(alnum=True)
1447            else:
1448                break
1449
1450        self._add(
1451            TokenType.VAR
1452            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1453            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1454        )
1455
1456    def _extract_string(
1457        self,
1458        delimiter: str,
1459        escapes: t.Optional[t.Set[str]] = None,
1460        raw_string: bool = False,
1461        raise_unmatched: bool = True,
1462    ) -> str:
1463        text = ""
1464        delim_size = len(delimiter)
1465        escapes = self._STRING_ESCAPES if escapes is None else escapes
1466
1467        while True:
1468            if (
1469                not raw_string
1470                and self.dialect.UNESCAPED_SEQUENCES
1471                and self._peek
1472                and self._char in self.STRING_ESCAPES
1473            ):
1474                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1475                if unescaped_sequence:
1476                    self._advance(2)
1477                    text += unescaped_sequence
1478                    continue
1479            if (
1480                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1481                and self._char in escapes
1482                and (self._peek == delimiter or self._peek in escapes)
1483                and (self._char not in self._QUOTES or self._char == self._peek)
1484            ):
1485                if self._peek == delimiter:
1486                    text += self._peek
1487                else:
1488                    text += self._char + self._peek
1489
1490                if self._current + 1 < self.size:
1491                    self._advance(2)
1492                else:
1493                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1494            else:
1495                if self._chars(delim_size) == delimiter:
1496                    if delim_size > 1:
1497                        self._advance(delim_size - 1)
1498                    break
1499
1500                if self._end:
1501                    if not raise_unmatched:
1502                        return text + self._char
1503
1504                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1505
1506                current = self._current - 1
1507                self._advance(alnum=True)
1508                text += self.sql[current : self._current - 1]
1509
1510        return text
1511
1512    def tokenize_rs(self, sql: str) -> t.List[Token]:
1513        if not self._RS_TOKENIZER:
1514            raise SqlglotError("Rust tokenizer is not available")
1515
1516        tokens, error_msg = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1517        for token in tokens:
1518            token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1519
1520        # Setting this here so partial token lists can be inspected even if there is a failure
1521        self.tokens = tokens
1522
1523        if error_msg is not None:
1524            raise TokenError(error_msg)
1525
1526        return tokens