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    SEMANTIC_VIEW = auto()
 416    VOLATILE = auto()
 417    WHEN = auto()
 418    WHERE = auto()
 419    WINDOW = auto()
 420    WITH = auto()
 421    UNIQUE = auto()
 422    VERSION_SNAPSHOT = auto()
 423    TIMESTAMP_SNAPSHOT = auto()
 424    OPTION = auto()
 425    SINK = auto()
 426    SOURCE = auto()
 427    ANALYZE = auto()
 428    NAMESPACE = auto()
 429    EXPORT = auto()
 430
 431    # sentinel
 432    HIVE_TOKEN_STREAM = auto()
 433
 434
 435_ALL_TOKEN_TYPES = list(TokenType)
 436_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 437
 438
 439class Token:
 440    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 441
 442    @classmethod
 443    def number(cls, number: int) -> Token:
 444        """Returns a NUMBER token with `number` as its text."""
 445        return cls(TokenType.NUMBER, str(number))
 446
 447    @classmethod
 448    def string(cls, string: str) -> Token:
 449        """Returns a STRING token with `string` as its text."""
 450        return cls(TokenType.STRING, string)
 451
 452    @classmethod
 453    def identifier(cls, identifier: str) -> Token:
 454        """Returns an IDENTIFIER token with `identifier` as its text."""
 455        return cls(TokenType.IDENTIFIER, identifier)
 456
 457    @classmethod
 458    def var(cls, var: str) -> Token:
 459        """Returns an VAR token with `var` as its text."""
 460        return cls(TokenType.VAR, var)
 461
 462    def __init__(
 463        self,
 464        token_type: TokenType,
 465        text: str,
 466        line: int = 1,
 467        col: int = 1,
 468        start: int = 0,
 469        end: int = 0,
 470        comments: t.Optional[t.List[str]] = None,
 471    ) -> None:
 472        """Token initializer.
 473
 474        Args:
 475            token_type: The TokenType Enum.
 476            text: The text of the token.
 477            line: The line that the token ends on.
 478            col: The column that the token ends on.
 479            start: The start index of the token.
 480            end: The ending index of the token.
 481            comments: The comments to attach to the token.
 482        """
 483        self.token_type = token_type
 484        self.text = text
 485        self.line = line
 486        self.col = col
 487        self.start = start
 488        self.end = end
 489        self.comments = [] if comments is None else comments
 490
 491    def __repr__(self) -> str:
 492        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 493        return f"<Token {attributes}>"
 494
 495
 496class _Tokenizer(type):
 497    def __new__(cls, clsname, bases, attrs):
 498        klass = super().__new__(cls, clsname, bases, attrs)
 499
 500        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 501            return dict(
 502                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 503            )
 504
 505        def _quotes_to_format(
 506            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 507        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 508            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 509
 510        klass._QUOTES = _convert_quotes(klass.QUOTES)
 511        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 512
 513        klass._FORMAT_STRINGS = {
 514            **{
 515                p + s: (e, TokenType.NATIONAL_STRING)
 516                for s, e in klass._QUOTES.items()
 517                for p in ("n", "N")
 518            },
 519            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 520            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 521            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 522            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 523            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 524            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 525        }
 526
 527        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 528        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 529        klass._COMMENTS = {
 530            **dict(
 531                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 532                for comment in klass.COMMENTS
 533            ),
 534            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 535        }
 536        if klass.HINT_START in klass.KEYWORDS:
 537            klass._COMMENTS[klass.HINT_START] = "*/"
 538
 539        klass._KEYWORD_TRIE = new_trie(
 540            key.upper()
 541            for key in (
 542                *klass.KEYWORDS,
 543                *klass._COMMENTS,
 544                *klass._QUOTES,
 545                *klass._FORMAT_STRINGS,
 546            )
 547            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 548        )
 549
 550        if USE_RS_TOKENIZER:
 551            settings = RsTokenizerSettings(
 552                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 553                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 554                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 555                numeric_literals=klass.NUMERIC_LITERALS,
 556                identifiers=klass._IDENTIFIERS,
 557                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 558                string_escapes=klass._STRING_ESCAPES,
 559                quotes=klass._QUOTES,
 560                format_strings={
 561                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 562                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 563                },
 564                has_bit_strings=bool(klass.BIT_STRINGS),
 565                has_hex_strings=bool(klass.HEX_STRINGS),
 566                comments=klass._COMMENTS,
 567                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 568                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 569                command_prefix_tokens={
 570                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 571                },
 572                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 573                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 574                nested_comments=klass.NESTED_COMMENTS,
 575                hint_start=klass.HINT_START,
 576                tokens_preceding_hint={
 577                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 578                },
 579            )
 580            token_types = RsTokenTypeSettings(
 581                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 582                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 583                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 584                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 585                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 586                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 587                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 588                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 589                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 590                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 591                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 592                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 593                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 594                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 595            )
 596            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 597        else:
 598            klass._RS_TOKENIZER = None
 599
 600        return klass
 601
 602
 603class Tokenizer(metaclass=_Tokenizer):
 604    SINGLE_TOKENS = {
 605        "(": TokenType.L_PAREN,
 606        ")": TokenType.R_PAREN,
 607        "[": TokenType.L_BRACKET,
 608        "]": TokenType.R_BRACKET,
 609        "{": TokenType.L_BRACE,
 610        "}": TokenType.R_BRACE,
 611        "&": TokenType.AMP,
 612        "^": TokenType.CARET,
 613        ":": TokenType.COLON,
 614        ",": TokenType.COMMA,
 615        ".": TokenType.DOT,
 616        "-": TokenType.DASH,
 617        "=": TokenType.EQ,
 618        ">": TokenType.GT,
 619        "<": TokenType.LT,
 620        "%": TokenType.MOD,
 621        "!": TokenType.NOT,
 622        "|": TokenType.PIPE,
 623        "+": TokenType.PLUS,
 624        ";": TokenType.SEMICOLON,
 625        "/": TokenType.SLASH,
 626        "\\": TokenType.BACKSLASH,
 627        "*": TokenType.STAR,
 628        "~": TokenType.TILDA,
 629        "?": TokenType.PLACEHOLDER,
 630        "@": TokenType.PARAMETER,
 631        "#": TokenType.HASH,
 632        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 633        "'": TokenType.UNKNOWN,
 634        "`": TokenType.UNKNOWN,
 635        '"': TokenType.UNKNOWN,
 636    }
 637
 638    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 639    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 640    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 641    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 642    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 643    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 644    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 645    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 646    STRING_ESCAPES = ["'"]
 647    VAR_SINGLE_TOKENS: t.Set[str] = set()
 648
 649    # The strings in this list can always be used as escapes, regardless of the surrounding
 650    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 651    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 652    IDENTIFIER_ESCAPES: t.List[str] = []
 653
 654    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 655    HEREDOC_TAG_IS_IDENTIFIER = False
 656
 657    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 658    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 659
 660    # Whether string escape characters function as such when placed within raw strings
 661    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 662
 663    NESTED_COMMENTS = True
 664
 665    HINT_START = "/*+"
 666
 667    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 668
 669    # Autofilled
 670    _COMMENTS: t.Dict[str, str] = {}
 671    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 672    _IDENTIFIERS: t.Dict[str, str] = {}
 673    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 674    _QUOTES: t.Dict[str, str] = {}
 675    _STRING_ESCAPES: t.Set[str] = set()
 676    _KEYWORD_TRIE: t.Dict = {}
 677    _RS_TOKENIZER: t.Optional[t.Any] = None
 678
 679    KEYWORDS: t.Dict[str, TokenType] = {
 680        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 681        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 682        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 683        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 684        HINT_START: TokenType.HINT,
 685        "==": TokenType.EQ,
 686        "::": TokenType.DCOLON,
 687        "||": TokenType.DPIPE,
 688        "|>": TokenType.PIPE_GT,
 689        ">=": TokenType.GTE,
 690        "<=": TokenType.LTE,
 691        "<>": TokenType.NEQ,
 692        "!=": TokenType.NEQ,
 693        ":=": TokenType.COLON_EQ,
 694        "<=>": TokenType.NULLSAFE_EQ,
 695        "->": TokenType.ARROW,
 696        "->>": TokenType.DARROW,
 697        "=>": TokenType.FARROW,
 698        "#>": TokenType.HASH_ARROW,
 699        "#>>": TokenType.DHASH_ARROW,
 700        "<->": TokenType.LR_ARROW,
 701        "&&": TokenType.DAMP,
 702        "??": TokenType.DQMARK,
 703        "~~~": TokenType.GLOB,
 704        "~~": TokenType.LIKE,
 705        "~~*": TokenType.ILIKE,
 706        "~*": TokenType.IRLIKE,
 707        "ALL": TokenType.ALL,
 708        "ALWAYS": TokenType.ALWAYS,
 709        "AND": TokenType.AND,
 710        "ANTI": TokenType.ANTI,
 711        "ANY": TokenType.ANY,
 712        "ASC": TokenType.ASC,
 713        "AS": TokenType.ALIAS,
 714        "ASOF": TokenType.ASOF,
 715        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 716        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 717        "BEGIN": TokenType.BEGIN,
 718        "BETWEEN": TokenType.BETWEEN,
 719        "CACHE": TokenType.CACHE,
 720        "UNCACHE": TokenType.UNCACHE,
 721        "CASE": TokenType.CASE,
 722        "CHARACTER SET": TokenType.CHARACTER_SET,
 723        "CLUSTER BY": TokenType.CLUSTER_BY,
 724        "COLLATE": TokenType.COLLATE,
 725        "COLUMN": TokenType.COLUMN,
 726        "COMMIT": TokenType.COMMIT,
 727        "CONNECT BY": TokenType.CONNECT_BY,
 728        "CONSTRAINT": TokenType.CONSTRAINT,
 729        "COPY": TokenType.COPY,
 730        "CREATE": TokenType.CREATE,
 731        "CROSS": TokenType.CROSS,
 732        "CUBE": TokenType.CUBE,
 733        "CURRENT_DATE": TokenType.CURRENT_DATE,
 734        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 735        "CURRENT_TIME": TokenType.CURRENT_TIME,
 736        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 737        "CURRENT_USER": TokenType.CURRENT_USER,
 738        "DATABASE": TokenType.DATABASE,
 739        "DEFAULT": TokenType.DEFAULT,
 740        "DELETE": TokenType.DELETE,
 741        "DESC": TokenType.DESC,
 742        "DESCRIBE": TokenType.DESCRIBE,
 743        "DISTINCT": TokenType.DISTINCT,
 744        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 745        "DIV": TokenType.DIV,
 746        "DROP": TokenType.DROP,
 747        "ELSE": TokenType.ELSE,
 748        "END": TokenType.END,
 749        "ENUM": TokenType.ENUM,
 750        "ESCAPE": TokenType.ESCAPE,
 751        "EXCEPT": TokenType.EXCEPT,
 752        "EXECUTE": TokenType.EXECUTE,
 753        "EXISTS": TokenType.EXISTS,
 754        "FALSE": TokenType.FALSE,
 755        "FETCH": TokenType.FETCH,
 756        "FILTER": TokenType.FILTER,
 757        "FIRST": TokenType.FIRST,
 758        "FULL": TokenType.FULL,
 759        "FUNCTION": TokenType.FUNCTION,
 760        "FOR": TokenType.FOR,
 761        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 762        "FORMAT": TokenType.FORMAT,
 763        "FROM": TokenType.FROM,
 764        "GEOGRAPHY": TokenType.GEOGRAPHY,
 765        "GEOMETRY": TokenType.GEOMETRY,
 766        "GLOB": TokenType.GLOB,
 767        "GROUP BY": TokenType.GROUP_BY,
 768        "GROUPING SETS": TokenType.GROUPING_SETS,
 769        "HAVING": TokenType.HAVING,
 770        "ILIKE": TokenType.ILIKE,
 771        "IN": TokenType.IN,
 772        "INDEX": TokenType.INDEX,
 773        "INET": TokenType.INET,
 774        "INNER": TokenType.INNER,
 775        "INSERT": TokenType.INSERT,
 776        "INTERVAL": TokenType.INTERVAL,
 777        "INTERSECT": TokenType.INTERSECT,
 778        "INTO": TokenType.INTO,
 779        "IS": TokenType.IS,
 780        "ISNULL": TokenType.ISNULL,
 781        "JOIN": TokenType.JOIN,
 782        "KEEP": TokenType.KEEP,
 783        "KILL": TokenType.KILL,
 784        "LATERAL": TokenType.LATERAL,
 785        "LEFT": TokenType.LEFT,
 786        "LIKE": TokenType.LIKE,
 787        "LIMIT": TokenType.LIMIT,
 788        "LOAD": TokenType.LOAD,
 789        "LOCK": TokenType.LOCK,
 790        "MERGE": TokenType.MERGE,
 791        "NAMESPACE": TokenType.NAMESPACE,
 792        "NATURAL": TokenType.NATURAL,
 793        "NEXT": TokenType.NEXT,
 794        "NOT": TokenType.NOT,
 795        "NOTNULL": TokenType.NOTNULL,
 796        "NULL": TokenType.NULL,
 797        "OBJECT": TokenType.OBJECT,
 798        "OFFSET": TokenType.OFFSET,
 799        "ON": TokenType.ON,
 800        "OR": TokenType.OR,
 801        "XOR": TokenType.XOR,
 802        "ORDER BY": TokenType.ORDER_BY,
 803        "ORDINALITY": TokenType.ORDINALITY,
 804        "OUTER": TokenType.OUTER,
 805        "OVER": TokenType.OVER,
 806        "OVERLAPS": TokenType.OVERLAPS,
 807        "OVERWRITE": TokenType.OVERWRITE,
 808        "PARTITION": TokenType.PARTITION,
 809        "PARTITION BY": TokenType.PARTITION_BY,
 810        "PARTITIONED BY": TokenType.PARTITION_BY,
 811        "PARTITIONED_BY": TokenType.PARTITION_BY,
 812        "PERCENT": TokenType.PERCENT,
 813        "PIVOT": TokenType.PIVOT,
 814        "PRAGMA": TokenType.PRAGMA,
 815        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 816        "PROCEDURE": TokenType.PROCEDURE,
 817        "QUALIFY": TokenType.QUALIFY,
 818        "RANGE": TokenType.RANGE,
 819        "RECURSIVE": TokenType.RECURSIVE,
 820        "REGEXP": TokenType.RLIKE,
 821        "RENAME": TokenType.RENAME,
 822        "REPLACE": TokenType.REPLACE,
 823        "RETURNING": TokenType.RETURNING,
 824        "REFERENCES": TokenType.REFERENCES,
 825        "RIGHT": TokenType.RIGHT,
 826        "RLIKE": TokenType.RLIKE,
 827        "ROLLBACK": TokenType.ROLLBACK,
 828        "ROLLUP": TokenType.ROLLUP,
 829        "ROW": TokenType.ROW,
 830        "ROWS": TokenType.ROWS,
 831        "SCHEMA": TokenType.SCHEMA,
 832        "SELECT": TokenType.SELECT,
 833        "SEMI": TokenType.SEMI,
 834        "SET": TokenType.SET,
 835        "SETTINGS": TokenType.SETTINGS,
 836        "SHOW": TokenType.SHOW,
 837        "SIMILAR TO": TokenType.SIMILAR_TO,
 838        "SOME": TokenType.SOME,
 839        "SORT BY": TokenType.SORT_BY,
 840        "START WITH": TokenType.START_WITH,
 841        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 842        "TABLE": TokenType.TABLE,
 843        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 844        "TEMP": TokenType.TEMPORARY,
 845        "TEMPORARY": TokenType.TEMPORARY,
 846        "THEN": TokenType.THEN,
 847        "TRUE": TokenType.TRUE,
 848        "TRUNCATE": TokenType.TRUNCATE,
 849        "UNION": TokenType.UNION,
 850        "UNKNOWN": TokenType.UNKNOWN,
 851        "UNNEST": TokenType.UNNEST,
 852        "UNPIVOT": TokenType.UNPIVOT,
 853        "UPDATE": TokenType.UPDATE,
 854        "USE": TokenType.USE,
 855        "USING": TokenType.USING,
 856        "UUID": TokenType.UUID,
 857        "VALUES": TokenType.VALUES,
 858        "VIEW": TokenType.VIEW,
 859        "VOLATILE": TokenType.VOLATILE,
 860        "WHEN": TokenType.WHEN,
 861        "WHERE": TokenType.WHERE,
 862        "WINDOW": TokenType.WINDOW,
 863        "WITH": TokenType.WITH,
 864        "APPLY": TokenType.APPLY,
 865        "ARRAY": TokenType.ARRAY,
 866        "BIT": TokenType.BIT,
 867        "BOOL": TokenType.BOOLEAN,
 868        "BOOLEAN": TokenType.BOOLEAN,
 869        "BYTE": TokenType.TINYINT,
 870        "MEDIUMINT": TokenType.MEDIUMINT,
 871        "INT1": TokenType.TINYINT,
 872        "TINYINT": TokenType.TINYINT,
 873        "INT16": TokenType.SMALLINT,
 874        "SHORT": TokenType.SMALLINT,
 875        "SMALLINT": TokenType.SMALLINT,
 876        "HUGEINT": TokenType.INT128,
 877        "UHUGEINT": TokenType.UINT128,
 878        "INT2": TokenType.SMALLINT,
 879        "INTEGER": TokenType.INT,
 880        "INT": TokenType.INT,
 881        "INT4": TokenType.INT,
 882        "INT32": TokenType.INT,
 883        "INT64": TokenType.BIGINT,
 884        "INT128": TokenType.INT128,
 885        "INT256": TokenType.INT256,
 886        "LONG": TokenType.BIGINT,
 887        "BIGINT": TokenType.BIGINT,
 888        "INT8": TokenType.TINYINT,
 889        "UINT": TokenType.UINT,
 890        "UINT128": TokenType.UINT128,
 891        "UINT256": TokenType.UINT256,
 892        "DEC": TokenType.DECIMAL,
 893        "DECIMAL": TokenType.DECIMAL,
 894        "DECIMAL32": TokenType.DECIMAL32,
 895        "DECIMAL64": TokenType.DECIMAL64,
 896        "DECIMAL128": TokenType.DECIMAL128,
 897        "DECIMAL256": TokenType.DECIMAL256,
 898        "BIGDECIMAL": TokenType.BIGDECIMAL,
 899        "BIGNUMERIC": TokenType.BIGDECIMAL,
 900        "LIST": TokenType.LIST,
 901        "MAP": TokenType.MAP,
 902        "NULLABLE": TokenType.NULLABLE,
 903        "NUMBER": TokenType.DECIMAL,
 904        "NUMERIC": TokenType.DECIMAL,
 905        "FIXED": TokenType.DECIMAL,
 906        "REAL": TokenType.FLOAT,
 907        "FLOAT": TokenType.FLOAT,
 908        "FLOAT4": TokenType.FLOAT,
 909        "FLOAT8": TokenType.DOUBLE,
 910        "DOUBLE": TokenType.DOUBLE,
 911        "DOUBLE PRECISION": TokenType.DOUBLE,
 912        "JSON": TokenType.JSON,
 913        "JSONB": TokenType.JSONB,
 914        "CHAR": TokenType.CHAR,
 915        "CHARACTER": TokenType.CHAR,
 916        "CHAR VARYING": TokenType.VARCHAR,
 917        "CHARACTER VARYING": TokenType.VARCHAR,
 918        "NCHAR": TokenType.NCHAR,
 919        "VARCHAR": TokenType.VARCHAR,
 920        "VARCHAR2": TokenType.VARCHAR,
 921        "NVARCHAR": TokenType.NVARCHAR,
 922        "NVARCHAR2": TokenType.NVARCHAR,
 923        "BPCHAR": TokenType.BPCHAR,
 924        "STR": TokenType.TEXT,
 925        "STRING": TokenType.TEXT,
 926        "TEXT": TokenType.TEXT,
 927        "LONGTEXT": TokenType.LONGTEXT,
 928        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 929        "TINYTEXT": TokenType.TINYTEXT,
 930        "CLOB": TokenType.TEXT,
 931        "LONGVARCHAR": TokenType.TEXT,
 932        "BINARY": TokenType.BINARY,
 933        "BLOB": TokenType.VARBINARY,
 934        "LONGBLOB": TokenType.LONGBLOB,
 935        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 936        "TINYBLOB": TokenType.TINYBLOB,
 937        "BYTEA": TokenType.VARBINARY,
 938        "VARBINARY": TokenType.VARBINARY,
 939        "TIME": TokenType.TIME,
 940        "TIMETZ": TokenType.TIMETZ,
 941        "TIMESTAMP": TokenType.TIMESTAMP,
 942        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 943        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 944        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 945        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 946        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 947        "DATE": TokenType.DATE,
 948        "DATETIME": TokenType.DATETIME,
 949        "INT4RANGE": TokenType.INT4RANGE,
 950        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 951        "INT8RANGE": TokenType.INT8RANGE,
 952        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 953        "NUMRANGE": TokenType.NUMRANGE,
 954        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 955        "TSRANGE": TokenType.TSRANGE,
 956        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 957        "TSTZRANGE": TokenType.TSTZRANGE,
 958        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 959        "DATERANGE": TokenType.DATERANGE,
 960        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 961        "UNIQUE": TokenType.UNIQUE,
 962        "VECTOR": TokenType.VECTOR,
 963        "STRUCT": TokenType.STRUCT,
 964        "SEQUENCE": TokenType.SEQUENCE,
 965        "VARIANT": TokenType.VARIANT,
 966        "ALTER": TokenType.ALTER,
 967        "ANALYZE": TokenType.ANALYZE,
 968        "CALL": TokenType.COMMAND,
 969        "COMMENT": TokenType.COMMENT,
 970        "EXPLAIN": TokenType.COMMAND,
 971        "GRANT": TokenType.GRANT,
 972        "OPTIMIZE": TokenType.COMMAND,
 973        "PREPARE": TokenType.COMMAND,
 974        "VACUUM": TokenType.COMMAND,
 975        "USER-DEFINED": TokenType.USERDEFINED,
 976        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 977        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 978    }
 979
 980    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 981        " ": TokenType.SPACE,
 982        "\t": TokenType.SPACE,
 983        "\n": TokenType.BREAK,
 984        "\r": TokenType.BREAK,
 985    }
 986
 987    COMMANDS = {
 988        TokenType.COMMAND,
 989        TokenType.EXECUTE,
 990        TokenType.FETCH,
 991        TokenType.SHOW,
 992        TokenType.RENAME,
 993    }
 994
 995    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 996
 997    # Handle numeric literals like in hive (3L = BIGINT)
 998    NUMERIC_LITERALS: t.Dict[str, str] = {}
 999
1000    COMMENTS = ["--", ("/*", "*/")]
1001
1002    __slots__ = (
1003        "sql",
1004        "size",
1005        "tokens",
1006        "dialect",
1007        "use_rs_tokenizer",
1008        "_start",
1009        "_current",
1010        "_line",
1011        "_col",
1012        "_comments",
1013        "_char",
1014        "_end",
1015        "_peek",
1016        "_prev_token_line",
1017        "_rs_dialect_settings",
1018    )
1019
1020    def __init__(
1021        self,
1022        dialect: DialectType = None,
1023        use_rs_tokenizer: t.Optional[bool] = None,
1024        **opts: t.Any,
1025    ) -> None:
1026        from sqlglot.dialects import Dialect
1027
1028        self.dialect = Dialect.get_or_raise(dialect)
1029
1030        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1031        self.use_rs_tokenizer = (
1032            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1033        )
1034
1035        if self.use_rs_tokenizer:
1036            self._rs_dialect_settings = RsTokenizerDialectSettings(
1037                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1038                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1039                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1040            )
1041
1042        self.reset()
1043
1044    def reset(self) -> None:
1045        self.sql = ""
1046        self.size = 0
1047        self.tokens: t.List[Token] = []
1048        self._start = 0
1049        self._current = 0
1050        self._line = 1
1051        self._col = 0
1052        self._comments: t.List[str] = []
1053
1054        self._char = ""
1055        self._end = False
1056        self._peek = ""
1057        self._prev_token_line = -1
1058
1059    def tokenize(self, sql: str) -> t.List[Token]:
1060        """Returns a list of tokens corresponding to the SQL string `sql`."""
1061        if self.use_rs_tokenizer:
1062            return self.tokenize_rs(sql)
1063
1064        self.reset()
1065        self.sql = sql
1066        self.size = len(sql)
1067
1068        try:
1069            self._scan()
1070        except Exception as e:
1071            start = max(self._current - 50, 0)
1072            end = min(self._current + 50, self.size - 1)
1073            context = self.sql[start:end]
1074            raise TokenError(f"Error tokenizing '{context}'") from e
1075
1076        return self.tokens
1077
1078    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1079        while self.size and not self._end:
1080            current = self._current
1081
1082            # Skip spaces here rather than iteratively calling advance() for performance reasons
1083            while current < self.size:
1084                char = self.sql[current]
1085
1086                if char.isspace() and (char == " " or char == "\t"):
1087                    current += 1
1088                else:
1089                    break
1090
1091            offset = current - self._current if current > self._current else 1
1092
1093            self._start = current
1094            self._advance(offset)
1095
1096            if not self._char.isspace():
1097                if self._char.isdigit():
1098                    self._scan_number()
1099                elif self._char in self._IDENTIFIERS:
1100                    self._scan_identifier(self._IDENTIFIERS[self._char])
1101                else:
1102                    self._scan_keywords()
1103
1104            if until and until():
1105                break
1106
1107        if self.tokens and self._comments:
1108            self.tokens[-1].comments.extend(self._comments)
1109
1110    def _chars(self, size: int) -> str:
1111        if size == 1:
1112            return self._char
1113
1114        start = self._current - 1
1115        end = start + size
1116
1117        return self.sql[start:end] if end <= self.size else ""
1118
1119    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1120        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1121            # Ensures we don't count an extra line if we get a \r\n line break sequence
1122            if not (self._char == "\r" and self._peek == "\n"):
1123                self._col = i
1124                self._line += 1
1125        else:
1126            self._col += i
1127
1128        self._current += i
1129        self._end = self._current >= self.size
1130        self._char = self.sql[self._current - 1]
1131        self._peek = "" if self._end else self.sql[self._current]
1132
1133        if alnum and self._char.isalnum():
1134            # Here we use local variables instead of attributes for better performance
1135            _col = self._col
1136            _current = self._current
1137            _end = self._end
1138            _peek = self._peek
1139
1140            while _peek.isalnum():
1141                _col += 1
1142                _current += 1
1143                _end = _current >= self.size
1144                _peek = "" if _end else self.sql[_current]
1145
1146            self._col = _col
1147            self._current = _current
1148            self._end = _end
1149            self._peek = _peek
1150            self._char = self.sql[_current - 1]
1151
1152    @property
1153    def _text(self) -> str:
1154        return self.sql[self._start : self._current]
1155
1156    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1157        self._prev_token_line = self._line
1158
1159        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1160            self.tokens[-1].comments.extend(self._comments)
1161            self._comments = []
1162
1163        self.tokens.append(
1164            Token(
1165                token_type,
1166                text=self._text if text is None else text,
1167                line=self._line,
1168                col=self._col,
1169                start=self._start,
1170                end=self._current - 1,
1171                comments=self._comments,
1172            )
1173        )
1174        self._comments = []
1175
1176        # If we have either a semicolon or a begin token before the command's token, we'll parse
1177        # whatever follows the command's token as a string
1178        if (
1179            token_type in self.COMMANDS
1180            and self._peek != ";"
1181            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1182        ):
1183            start = self._current
1184            tokens = len(self.tokens)
1185            self._scan(lambda: self._peek == ";")
1186            self.tokens = self.tokens[:tokens]
1187            text = self.sql[start : self._current].strip()
1188            if text:
1189                self._add(TokenType.STRING, text)
1190
1191    def _scan_keywords(self) -> None:
1192        size = 0
1193        word = None
1194        chars = self._text
1195        char = chars
1196        prev_space = False
1197        skip = False
1198        trie = self._KEYWORD_TRIE
1199        single_token = char in self.SINGLE_TOKENS
1200
1201        while chars:
1202            if skip:
1203                result = TrieResult.PREFIX
1204            else:
1205                result, trie = in_trie(trie, char.upper())
1206
1207            if result == TrieResult.FAILED:
1208                break
1209            if result == TrieResult.EXISTS:
1210                word = chars
1211
1212            end = self._current + size
1213            size += 1
1214
1215            if end < self.size:
1216                char = self.sql[end]
1217                single_token = single_token or char in self.SINGLE_TOKENS
1218                is_space = char.isspace()
1219
1220                if not is_space or not prev_space:
1221                    if is_space:
1222                        char = " "
1223                    chars += char
1224                    prev_space = is_space
1225                    skip = False
1226                else:
1227                    skip = True
1228            else:
1229                char = ""
1230                break
1231
1232        if word:
1233            if self._scan_string(word):
1234                return
1235            if self._scan_comment(word):
1236                return
1237            if prev_space or single_token or not char:
1238                self._advance(size - 1)
1239                word = word.upper()
1240                self._add(self.KEYWORDS[word], text=word)
1241                return
1242
1243        if self._char in self.SINGLE_TOKENS:
1244            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1245            return
1246
1247        self._scan_var()
1248
1249    def _scan_comment(self, comment_start: str) -> bool:
1250        if comment_start not in self._COMMENTS:
1251            return False
1252
1253        comment_start_line = self._line
1254        comment_start_size = len(comment_start)
1255        comment_end = self._COMMENTS[comment_start]
1256
1257        if comment_end:
1258            # Skip the comment's start delimiter
1259            self._advance(comment_start_size)
1260
1261            comment_count = 1
1262            comment_end_size = len(comment_end)
1263
1264            while not self._end:
1265                if self._chars(comment_end_size) == comment_end:
1266                    comment_count -= 1
1267                    if not comment_count:
1268                        break
1269
1270                self._advance(alnum=True)
1271
1272                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1273                if (
1274                    self.NESTED_COMMENTS
1275                    and not self._end
1276                    and self._chars(comment_end_size) == comment_start
1277                ):
1278                    self._advance(comment_start_size)
1279                    comment_count += 1
1280
1281            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1282            self._advance(comment_end_size - 1)
1283        else:
1284            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1285                self._advance(alnum=True)
1286            self._comments.append(self._text[comment_start_size:])
1287
1288        if (
1289            comment_start == self.HINT_START
1290            and self.tokens
1291            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1292        ):
1293            self._add(TokenType.HINT)
1294
1295        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1296        # Multiple consecutive comments are preserved by appending them to the current comments list.
1297        if comment_start_line == self._prev_token_line:
1298            self.tokens[-1].comments.extend(self._comments)
1299            self._comments = []
1300            self._prev_token_line = self._line
1301
1302        return True
1303
1304    def _scan_number(self) -> None:
1305        if self._char == "0":
1306            peek = self._peek.upper()
1307            if peek == "B":
1308                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1309            elif peek == "X":
1310                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1311
1312        decimal = False
1313        scientific = 0
1314
1315        while True:
1316            if self._peek.isdigit():
1317                self._advance()
1318            elif self._peek == "." and not decimal:
1319                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1320                    return self._add(TokenType.NUMBER)
1321                decimal = True
1322                self._advance()
1323            elif self._peek in ("-", "+") and scientific == 1:
1324                scientific += 1
1325                self._advance()
1326            elif self._peek.upper() == "E" and not scientific:
1327                scientific += 1
1328                self._advance()
1329            elif self._peek.isidentifier():
1330                number_text = self._text
1331                literal = ""
1332
1333                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1334                    literal += self._peek
1335                    self._advance()
1336
1337                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1338
1339                if token_type:
1340                    self._add(TokenType.NUMBER, number_text)
1341                    self._add(TokenType.DCOLON, "::")
1342                    return self._add(token_type, literal)
1343                else:
1344                    replaced = literal.replace("_", "")
1345                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1346                        return self._add(TokenType.NUMBER, number_text + replaced)
1347                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1348                        return self._add(TokenType.VAR)
1349
1350                self._advance(-len(literal))
1351                return self._add(TokenType.NUMBER, number_text)
1352            else:
1353                return self._add(TokenType.NUMBER)
1354
1355    def _scan_bits(self) -> None:
1356        self._advance()
1357        value = self._extract_value()
1358        try:
1359            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1360            int(value, 2)
1361            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1362        except ValueError:
1363            self._add(TokenType.IDENTIFIER)
1364
1365    def _scan_hex(self) -> None:
1366        self._advance()
1367        value = self._extract_value()
1368        try:
1369            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1370            int(value, 16)
1371            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1372        except ValueError:
1373            self._add(TokenType.IDENTIFIER)
1374
1375    def _extract_value(self) -> str:
1376        while True:
1377            char = self._peek.strip()
1378            if char and char not in self.SINGLE_TOKENS:
1379                self._advance(alnum=True)
1380            else:
1381                break
1382
1383        return self._text
1384
1385    def _scan_string(self, start: str) -> bool:
1386        base = None
1387        token_type = TokenType.STRING
1388
1389        if start in self._QUOTES:
1390            end = self._QUOTES[start]
1391        elif start in self._FORMAT_STRINGS:
1392            end, token_type = self._FORMAT_STRINGS[start]
1393
1394            if token_type == TokenType.HEX_STRING:
1395                base = 16
1396            elif token_type == TokenType.BIT_STRING:
1397                base = 2
1398            elif token_type == TokenType.HEREDOC_STRING:
1399                self._advance()
1400
1401                if self._char == end:
1402                    tag = ""
1403                else:
1404                    tag = self._extract_string(
1405                        end,
1406                        raw_string=True,
1407                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1408                    )
1409
1410                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1411                    if not self._end:
1412                        self._advance(-1)
1413
1414                    self._advance(-len(tag))
1415                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1416                    return True
1417
1418                end = f"{start}{tag}{end}"
1419        else:
1420            return False
1421
1422        self._advance(len(start))
1423        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1424
1425        if base:
1426            try:
1427                int(text, base)
1428            except Exception:
1429                raise TokenError(
1430                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1431                )
1432
1433        self._add(token_type, text)
1434        return True
1435
1436    def _scan_identifier(self, identifier_end: str) -> None:
1437        self._advance()
1438        text = self._extract_string(
1439            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1440        )
1441        self._add(TokenType.IDENTIFIER, text)
1442
1443    def _scan_var(self) -> None:
1444        while True:
1445            char = self._peek.strip()
1446            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1447                self._advance(alnum=True)
1448            else:
1449                break
1450
1451        self._add(
1452            TokenType.VAR
1453            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1454            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1455        )
1456
1457    def _extract_string(
1458        self,
1459        delimiter: str,
1460        escapes: t.Optional[t.Set[str]] = None,
1461        raw_string: bool = False,
1462        raise_unmatched: bool = True,
1463    ) -> str:
1464        text = ""
1465        delim_size = len(delimiter)
1466        escapes = self._STRING_ESCAPES if escapes is None else escapes
1467
1468        while True:
1469            if (
1470                not raw_string
1471                and self.dialect.UNESCAPED_SEQUENCES
1472                and self._peek
1473                and self._char in self.STRING_ESCAPES
1474            ):
1475                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1476                if unescaped_sequence:
1477                    self._advance(2)
1478                    text += unescaped_sequence
1479                    continue
1480            if (
1481                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1482                and self._char in escapes
1483                and (self._peek == delimiter or self._peek in escapes)
1484                and (self._char not in self._QUOTES or self._char == self._peek)
1485            ):
1486                if self._peek == delimiter:
1487                    text += self._peek
1488                else:
1489                    text += self._char + self._peek
1490
1491                if self._current + 1 < self.size:
1492                    self._advance(2)
1493                else:
1494                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1495            else:
1496                if self._chars(delim_size) == delimiter:
1497                    if delim_size > 1:
1498                        self._advance(delim_size - 1)
1499                    break
1500
1501                if self._end:
1502                    if not raise_unmatched:
1503                        return text + self._char
1504
1505                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1506
1507                current = self._current - 1
1508                self._advance(alnum=True)
1509                text += self.sql[current : self._current - 1]
1510
1511        return text
1512
1513    def tokenize_rs(self, sql: str) -> t.List[Token]:
1514        if not self._RS_TOKENIZER:
1515            raise SqlglotError("Rust tokenizer is not available")
1516
1517        tokens, error_msg = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1518        for token in tokens:
1519            token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1520
1521        # Setting this here so partial token lists can be inspected even if there is a failure
1522        self.tokens = tokens
1523
1524        if error_msg is not None:
1525            raise TokenError(error_msg)
1526
1527        return tokens