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    DCOLONDOLLAR = auto()
  43    DCOLONPERCENT = auto()
  44    DQMARK = auto()
  45    SEMICOLON = auto()
  46    STAR = auto()
  47    BACKSLASH = auto()
  48    SLASH = auto()
  49    LT = auto()
  50    LTE = auto()
  51    GT = auto()
  52    GTE = auto()
  53    NOT = auto()
  54    EQ = auto()
  55    NEQ = auto()
  56    NULLSAFE_EQ = auto()
  57    COLON_EQ = auto()
  58    COLON_GT = auto()
  59    NCOLON_GT = auto()
  60    AND = auto()
  61    OR = auto()
  62    AMP = auto()
  63    DPIPE = auto()
  64    PIPE_GT = auto()
  65    PIPE = auto()
  66    PIPE_SLASH = auto()
  67    DPIPE_SLASH = auto()
  68    CARET = auto()
  69    CARET_AT = auto()
  70    TILDA = auto()
  71    ARROW = auto()
  72    DARROW = auto()
  73    FARROW = auto()
  74    HASH = auto()
  75    HASH_ARROW = auto()
  76    DHASH_ARROW = auto()
  77    LR_ARROW = auto()
  78    DAT = auto()
  79    LT_AT = auto()
  80    AT_GT = auto()
  81    DOLLAR = auto()
  82    PARAMETER = auto()
  83    SESSION_PARAMETER = auto()
  84    DAMP = auto()
  85    XOR = auto()
  86    DSTAR = auto()
  87
  88    URI_START = auto()
  89
  90    BLOCK_START = auto()
  91    BLOCK_END = auto()
  92
  93    SPACE = auto()
  94    BREAK = auto()
  95
  96    STRING = auto()
  97    NUMBER = auto()
  98    IDENTIFIER = auto()
  99    DATABASE = auto()
 100    COLUMN = auto()
 101    COLUMN_DEF = auto()
 102    SCHEMA = auto()
 103    TABLE = auto()
 104    WAREHOUSE = auto()
 105    STAGE = auto()
 106    STREAMLIT = auto()
 107    VAR = auto()
 108    BIT_STRING = auto()
 109    HEX_STRING = auto()
 110    BYTE_STRING = auto()
 111    NATIONAL_STRING = auto()
 112    RAW_STRING = auto()
 113    HEREDOC_STRING = auto()
 114    UNICODE_STRING = auto()
 115
 116    # types
 117    BIT = auto()
 118    BOOLEAN = auto()
 119    TINYINT = auto()
 120    UTINYINT = auto()
 121    SMALLINT = auto()
 122    USMALLINT = auto()
 123    MEDIUMINT = auto()
 124    UMEDIUMINT = auto()
 125    INT = auto()
 126    UINT = auto()
 127    BIGINT = auto()
 128    UBIGINT = auto()
 129    INT128 = auto()
 130    UINT128 = auto()
 131    INT256 = auto()
 132    UINT256 = auto()
 133    FLOAT = auto()
 134    DOUBLE = auto()
 135    UDOUBLE = auto()
 136    DECIMAL = auto()
 137    DECIMAL32 = auto()
 138    DECIMAL64 = auto()
 139    DECIMAL128 = auto()
 140    DECIMAL256 = auto()
 141    UDECIMAL = auto()
 142    BIGDECIMAL = auto()
 143    CHAR = auto()
 144    NCHAR = auto()
 145    VARCHAR = auto()
 146    NVARCHAR = auto()
 147    BPCHAR = auto()
 148    TEXT = auto()
 149    MEDIUMTEXT = auto()
 150    LONGTEXT = auto()
 151    BLOB = auto()
 152    MEDIUMBLOB = auto()
 153    LONGBLOB = auto()
 154    TINYBLOB = auto()
 155    TINYTEXT = auto()
 156    NAME = auto()
 157    BINARY = auto()
 158    VARBINARY = auto()
 159    JSON = auto()
 160    JSONB = auto()
 161    TIME = auto()
 162    TIMETZ = auto()
 163    TIMESTAMP = auto()
 164    TIMESTAMPTZ = auto()
 165    TIMESTAMPLTZ = auto()
 166    TIMESTAMPNTZ = auto()
 167    TIMESTAMP_S = auto()
 168    TIMESTAMP_MS = auto()
 169    TIMESTAMP_NS = auto()
 170    DATETIME = auto()
 171    DATETIME2 = auto()
 172    DATETIME64 = auto()
 173    SMALLDATETIME = auto()
 174    DATE = auto()
 175    DATE32 = auto()
 176    INT4RANGE = auto()
 177    INT4MULTIRANGE = auto()
 178    INT8RANGE = auto()
 179    INT8MULTIRANGE = auto()
 180    NUMRANGE = auto()
 181    NUMMULTIRANGE = auto()
 182    TSRANGE = auto()
 183    TSMULTIRANGE = auto()
 184    TSTZRANGE = auto()
 185    TSTZMULTIRANGE = auto()
 186    DATERANGE = auto()
 187    DATEMULTIRANGE = auto()
 188    UUID = auto()
 189    GEOGRAPHY = auto()
 190    GEOGRAPHYPOINT = auto()
 191    NULLABLE = auto()
 192    GEOMETRY = auto()
 193    POINT = auto()
 194    RING = auto()
 195    LINESTRING = auto()
 196    MULTILINESTRING = auto()
 197    POLYGON = auto()
 198    MULTIPOLYGON = auto()
 199    HLLSKETCH = auto()
 200    HSTORE = auto()
 201    SUPER = auto()
 202    SERIAL = auto()
 203    SMALLSERIAL = auto()
 204    BIGSERIAL = auto()
 205    XML = auto()
 206    YEAR = auto()
 207    USERDEFINED = auto()
 208    MONEY = auto()
 209    SMALLMONEY = auto()
 210    ROWVERSION = auto()
 211    IMAGE = auto()
 212    VARIANT = auto()
 213    OBJECT = auto()
 214    INET = auto()
 215    IPADDRESS = auto()
 216    IPPREFIX = auto()
 217    IPV4 = auto()
 218    IPV6 = auto()
 219    ENUM = auto()
 220    ENUM8 = auto()
 221    ENUM16 = auto()
 222    FIXEDSTRING = auto()
 223    LOWCARDINALITY = auto()
 224    NESTED = auto()
 225    AGGREGATEFUNCTION = auto()
 226    SIMPLEAGGREGATEFUNCTION = auto()
 227    TDIGEST = auto()
 228    UNKNOWN = auto()
 229    VECTOR = auto()
 230    DYNAMIC = auto()
 231    VOID = auto()
 232
 233    # keywords
 234    ALIAS = auto()
 235    ALTER = auto()
 236    ALWAYS = auto()
 237    ALL = auto()
 238    ANTI = auto()
 239    ANY = auto()
 240    APPLY = auto()
 241    ARRAY = auto()
 242    ASC = auto()
 243    ASOF = auto()
 244    ATTACH = auto()
 245    AUTO_INCREMENT = auto()
 246    BEGIN = auto()
 247    BETWEEN = auto()
 248    BULK_COLLECT_INTO = auto()
 249    CACHE = auto()
 250    CASE = auto()
 251    CHARACTER_SET = auto()
 252    CLUSTER_BY = auto()
 253    COLLATE = auto()
 254    COMMAND = auto()
 255    COMMENT = auto()
 256    COMMIT = auto()
 257    CONNECT_BY = auto()
 258    CONSTRAINT = auto()
 259    COPY = auto()
 260    CREATE = auto()
 261    CROSS = auto()
 262    CUBE = auto()
 263    CURRENT_DATE = auto()
 264    CURRENT_DATETIME = auto()
 265    CURRENT_SCHEMA = auto()
 266    CURRENT_TIME = auto()
 267    CURRENT_TIMESTAMP = auto()
 268    CURRENT_USER = auto()
 269    DECLARE = auto()
 270    DEFAULT = auto()
 271    DELETE = auto()
 272    DESC = auto()
 273    DESCRIBE = auto()
 274    DETACH = auto()
 275    DICTIONARY = auto()
 276    DISTINCT = auto()
 277    DISTRIBUTE_BY = auto()
 278    DIV = auto()
 279    DROP = auto()
 280    ELSE = auto()
 281    END = auto()
 282    ESCAPE = auto()
 283    EXCEPT = auto()
 284    EXECUTE = auto()
 285    EXISTS = auto()
 286    FALSE = auto()
 287    FETCH = auto()
 288    FILE_FORMAT = auto()
 289    FILTER = auto()
 290    FINAL = auto()
 291    FIRST = auto()
 292    FOR = auto()
 293    FORCE = auto()
 294    FOREIGN_KEY = auto()
 295    FORMAT = auto()
 296    FROM = auto()
 297    FULL = auto()
 298    FUNCTION = auto()
 299    GET = auto()
 300    GLOB = auto()
 301    GLOBAL = auto()
 302    GRANT = auto()
 303    GROUP_BY = auto()
 304    GROUPING_SETS = auto()
 305    HAVING = auto()
 306    HINT = auto()
 307    IGNORE = auto()
 308    ILIKE = auto()
 309    IN = auto()
 310    INDEX = auto()
 311    INNER = auto()
 312    INSERT = auto()
 313    INTERSECT = auto()
 314    INTERVAL = auto()
 315    INTO = auto()
 316    INTRODUCER = auto()
 317    IRLIKE = auto()
 318    IS = auto()
 319    ISNULL = auto()
 320    JOIN = auto()
 321    JOIN_MARKER = auto()
 322    KEEP = auto()
 323    KEY = auto()
 324    KILL = auto()
 325    LANGUAGE = auto()
 326    LATERAL = auto()
 327    LEFT = auto()
 328    LIKE = auto()
 329    LIMIT = auto()
 330    LIST = auto()
 331    LOAD = auto()
 332    LOCK = auto()
 333    MAP = auto()
 334    MATCH_CONDITION = auto()
 335    MATCH_RECOGNIZE = auto()
 336    MEMBER_OF = auto()
 337    MERGE = auto()
 338    MOD = auto()
 339    MODEL = auto()
 340    NATURAL = auto()
 341    NEXT = auto()
 342    NOTHING = auto()
 343    NOTNULL = auto()
 344    NULL = auto()
 345    OBJECT_IDENTIFIER = auto()
 346    OFFSET = auto()
 347    ON = auto()
 348    ONLY = auto()
 349    OPERATOR = auto()
 350    ORDER_BY = auto()
 351    ORDER_SIBLINGS_BY = auto()
 352    ORDERED = auto()
 353    ORDINALITY = auto()
 354    OUTER = auto()
 355    OVER = auto()
 356    OVERLAPS = auto()
 357    OVERWRITE = auto()
 358    PARTITION = auto()
 359    PARTITION_BY = auto()
 360    PERCENT = auto()
 361    PIVOT = auto()
 362    PLACEHOLDER = auto()
 363    POSITIONAL = auto()
 364    PRAGMA = auto()
 365    PREWHERE = auto()
 366    PRIMARY_KEY = auto()
 367    PROCEDURE = auto()
 368    PROPERTIES = auto()
 369    PSEUDO_TYPE = auto()
 370    PUT = auto()
 371    QUALIFY = auto()
 372    QUOTE = auto()
 373    RANGE = auto()
 374    RECURSIVE = auto()
 375    REFRESH = auto()
 376    RENAME = auto()
 377    REPLACE = auto()
 378    RETURNING = auto()
 379    REFERENCES = auto()
 380    RIGHT = auto()
 381    RLIKE = auto()
 382    ROLLBACK = auto()
 383    ROLLUP = auto()
 384    ROW = auto()
 385    ROWS = auto()
 386    SELECT = auto()
 387    SEMI = auto()
 388    SEPARATOR = auto()
 389    SEQUENCE = auto()
 390    SERDE_PROPERTIES = auto()
 391    SET = auto()
 392    SETTINGS = auto()
 393    SHOW = auto()
 394    SIMILAR_TO = auto()
 395    SOME = auto()
 396    SORT_BY = auto()
 397    START_WITH = auto()
 398    STORAGE_INTEGRATION = auto()
 399    STRAIGHT_JOIN = auto()
 400    STRUCT = auto()
 401    SUMMARIZE = auto()
 402    TABLE_SAMPLE = auto()
 403    TAG = auto()
 404    TEMPORARY = auto()
 405    TOP = auto()
 406    THEN = auto()
 407    TRUE = auto()
 408    TRUNCATE = auto()
 409    UNCACHE = auto()
 410    UNION = auto()
 411    UNNEST = auto()
 412    UNPIVOT = auto()
 413    UPDATE = auto()
 414    USE = auto()
 415    USING = auto()
 416    VALUES = auto()
 417    VIEW = auto()
 418    SEMANTIC_VIEW = auto()
 419    VOLATILE = auto()
 420    WHEN = auto()
 421    WHERE = auto()
 422    WINDOW = auto()
 423    WITH = auto()
 424    UNIQUE = auto()
 425    VERSION_SNAPSHOT = auto()
 426    TIMESTAMP_SNAPSHOT = auto()
 427    OPTION = auto()
 428    SINK = auto()
 429    SOURCE = auto()
 430    ANALYZE = auto()
 431    NAMESPACE = auto()
 432    EXPORT = auto()
 433
 434    # sentinel
 435    HIVE_TOKEN_STREAM = auto()
 436
 437
 438_ALL_TOKEN_TYPES = list(TokenType)
 439_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 440
 441
 442class Token:
 443    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 444
 445    @classmethod
 446    def number(cls, number: int) -> Token:
 447        """Returns a NUMBER token with `number` as its text."""
 448        return cls(TokenType.NUMBER, str(number))
 449
 450    @classmethod
 451    def string(cls, string: str) -> Token:
 452        """Returns a STRING token with `string` as its text."""
 453        return cls(TokenType.STRING, string)
 454
 455    @classmethod
 456    def identifier(cls, identifier: str) -> Token:
 457        """Returns an IDENTIFIER token with `identifier` as its text."""
 458        return cls(TokenType.IDENTIFIER, identifier)
 459
 460    @classmethod
 461    def var(cls, var: str) -> Token:
 462        """Returns an VAR token with `var` as its text."""
 463        return cls(TokenType.VAR, var)
 464
 465    def __init__(
 466        self,
 467        token_type: TokenType,
 468        text: str,
 469        line: int = 1,
 470        col: int = 1,
 471        start: int = 0,
 472        end: int = 0,
 473        comments: t.Optional[t.List[str]] = None,
 474    ) -> None:
 475        """Token initializer.
 476
 477        Args:
 478            token_type: The TokenType Enum.
 479            text: The text of the token.
 480            line: The line that the token ends on.
 481            col: The column that the token ends on.
 482            start: The start index of the token.
 483            end: The ending index of the token.
 484            comments: The comments to attach to the token.
 485        """
 486        self.token_type = token_type
 487        self.text = text
 488        self.line = line
 489        self.col = col
 490        self.start = start
 491        self.end = end
 492        self.comments = [] if comments is None else comments
 493
 494    def __repr__(self) -> str:
 495        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 496        return f"<Token {attributes}>"
 497
 498
 499class _Tokenizer(type):
 500    def __new__(cls, clsname, bases, attrs):
 501        klass = super().__new__(cls, clsname, bases, attrs)
 502
 503        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 504            return dict(
 505                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 506            )
 507
 508        def _quotes_to_format(
 509            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 510        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 511            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 512
 513        klass._QUOTES = _convert_quotes(klass.QUOTES)
 514        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 515
 516        klass._FORMAT_STRINGS = {
 517            **{
 518                p + s: (e, TokenType.NATIONAL_STRING)
 519                for s, e in klass._QUOTES.items()
 520                for p in ("n", "N")
 521            },
 522            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 523            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 524            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 525            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 526            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 527            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 528        }
 529
 530        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 531        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 532        klass._COMMENTS = {
 533            **dict(
 534                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 535                for comment in klass.COMMENTS
 536            ),
 537            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 538        }
 539        if klass.HINT_START in klass.KEYWORDS:
 540            klass._COMMENTS[klass.HINT_START] = "*/"
 541
 542        klass._KEYWORD_TRIE = new_trie(
 543            key.upper()
 544            for key in (
 545                *klass.KEYWORDS,
 546                *klass._COMMENTS,
 547                *klass._QUOTES,
 548                *klass._FORMAT_STRINGS,
 549            )
 550            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 551        )
 552
 553        if USE_RS_TOKENIZER:
 554            settings = RsTokenizerSettings(
 555                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 556                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 557                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 558                numeric_literals=klass.NUMERIC_LITERALS,
 559                identifiers=klass._IDENTIFIERS,
 560                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 561                string_escapes=klass._STRING_ESCAPES,
 562                quotes=klass._QUOTES,
 563                format_strings={
 564                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 565                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 566                },
 567                has_bit_strings=bool(klass.BIT_STRINGS),
 568                has_hex_strings=bool(klass.HEX_STRINGS),
 569                comments=klass._COMMENTS,
 570                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 571                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 572                command_prefix_tokens={
 573                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 574                },
 575                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 576                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 577                nested_comments=klass.NESTED_COMMENTS,
 578                hint_start=klass.HINT_START,
 579                tokens_preceding_hint={
 580                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 581                },
 582            )
 583            token_types = RsTokenTypeSettings(
 584                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 585                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 586                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 587                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 588                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 589                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 590                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 591                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 592                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 593                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 594                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 595                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 596                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 597                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 598            )
 599            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 600        else:
 601            klass._RS_TOKENIZER = None
 602
 603        return klass
 604
 605
 606class Tokenizer(metaclass=_Tokenizer):
 607    SINGLE_TOKENS = {
 608        "(": TokenType.L_PAREN,
 609        ")": TokenType.R_PAREN,
 610        "[": TokenType.L_BRACKET,
 611        "]": TokenType.R_BRACKET,
 612        "{": TokenType.L_BRACE,
 613        "}": TokenType.R_BRACE,
 614        "&": TokenType.AMP,
 615        "^": TokenType.CARET,
 616        ":": TokenType.COLON,
 617        ",": TokenType.COMMA,
 618        ".": TokenType.DOT,
 619        "-": TokenType.DASH,
 620        "=": TokenType.EQ,
 621        ">": TokenType.GT,
 622        "<": TokenType.LT,
 623        "%": TokenType.MOD,
 624        "!": TokenType.NOT,
 625        "|": TokenType.PIPE,
 626        "+": TokenType.PLUS,
 627        ";": TokenType.SEMICOLON,
 628        "/": TokenType.SLASH,
 629        "\\": TokenType.BACKSLASH,
 630        "*": TokenType.STAR,
 631        "~": TokenType.TILDA,
 632        "?": TokenType.PLACEHOLDER,
 633        "@": TokenType.PARAMETER,
 634        "#": TokenType.HASH,
 635        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 636        "'": TokenType.UNKNOWN,
 637        "`": TokenType.UNKNOWN,
 638        '"': TokenType.UNKNOWN,
 639    }
 640
 641    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 642    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 643    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 644    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 645    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 646    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 647    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 648    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 649    STRING_ESCAPES = ["'"]
 650    VAR_SINGLE_TOKENS: t.Set[str] = set()
 651
 652    # The strings in this list can always be used as escapes, regardless of the surrounding
 653    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 654    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 655    IDENTIFIER_ESCAPES: t.List[str] = []
 656
 657    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 658    HEREDOC_TAG_IS_IDENTIFIER = False
 659
 660    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 661    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 662
 663    # Whether string escape characters function as such when placed within raw strings
 664    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 665
 666    NESTED_COMMENTS = True
 667
 668    HINT_START = "/*+"
 669
 670    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 671
 672    # Autofilled
 673    _COMMENTS: t.Dict[str, str] = {}
 674    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 675    _IDENTIFIERS: t.Dict[str, str] = {}
 676    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 677    _QUOTES: t.Dict[str, str] = {}
 678    _STRING_ESCAPES: t.Set[str] = set()
 679    _KEYWORD_TRIE: t.Dict = {}
 680    _RS_TOKENIZER: t.Optional[t.Any] = None
 681
 682    KEYWORDS: t.Dict[str, TokenType] = {
 683        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 684        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 685        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 686        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 687        HINT_START: TokenType.HINT,
 688        "==": TokenType.EQ,
 689        "::": TokenType.DCOLON,
 690        "||": TokenType.DPIPE,
 691        "|>": TokenType.PIPE_GT,
 692        ">=": TokenType.GTE,
 693        "<=": TokenType.LTE,
 694        "<>": TokenType.NEQ,
 695        "!=": TokenType.NEQ,
 696        ":=": TokenType.COLON_EQ,
 697        "<=>": TokenType.NULLSAFE_EQ,
 698        "->": TokenType.ARROW,
 699        "->>": TokenType.DARROW,
 700        "=>": TokenType.FARROW,
 701        "#>": TokenType.HASH_ARROW,
 702        "#>>": TokenType.DHASH_ARROW,
 703        "<->": TokenType.LR_ARROW,
 704        "&&": TokenType.DAMP,
 705        "??": TokenType.DQMARK,
 706        "~~~": TokenType.GLOB,
 707        "~~": TokenType.LIKE,
 708        "~~*": TokenType.ILIKE,
 709        "~*": TokenType.IRLIKE,
 710        "ALL": TokenType.ALL,
 711        "ALWAYS": TokenType.ALWAYS,
 712        "AND": TokenType.AND,
 713        "ANTI": TokenType.ANTI,
 714        "ANY": TokenType.ANY,
 715        "ASC": TokenType.ASC,
 716        "AS": TokenType.ALIAS,
 717        "ASOF": TokenType.ASOF,
 718        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 719        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 720        "BEGIN": TokenType.BEGIN,
 721        "BETWEEN": TokenType.BETWEEN,
 722        "CACHE": TokenType.CACHE,
 723        "UNCACHE": TokenType.UNCACHE,
 724        "CASE": TokenType.CASE,
 725        "CHARACTER SET": TokenType.CHARACTER_SET,
 726        "CLUSTER BY": TokenType.CLUSTER_BY,
 727        "COLLATE": TokenType.COLLATE,
 728        "COLUMN": TokenType.COLUMN,
 729        "COMMIT": TokenType.COMMIT,
 730        "CONNECT BY": TokenType.CONNECT_BY,
 731        "CONSTRAINT": TokenType.CONSTRAINT,
 732        "COPY": TokenType.COPY,
 733        "CREATE": TokenType.CREATE,
 734        "CROSS": TokenType.CROSS,
 735        "CUBE": TokenType.CUBE,
 736        "CURRENT_DATE": TokenType.CURRENT_DATE,
 737        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 738        "CURRENT_TIME": TokenType.CURRENT_TIME,
 739        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 740        "CURRENT_USER": TokenType.CURRENT_USER,
 741        "DATABASE": TokenType.DATABASE,
 742        "DEFAULT": TokenType.DEFAULT,
 743        "DELETE": TokenType.DELETE,
 744        "DESC": TokenType.DESC,
 745        "DESCRIBE": TokenType.DESCRIBE,
 746        "DISTINCT": TokenType.DISTINCT,
 747        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 748        "DIV": TokenType.DIV,
 749        "DROP": TokenType.DROP,
 750        "ELSE": TokenType.ELSE,
 751        "END": TokenType.END,
 752        "ENUM": TokenType.ENUM,
 753        "ESCAPE": TokenType.ESCAPE,
 754        "EXCEPT": TokenType.EXCEPT,
 755        "EXECUTE": TokenType.EXECUTE,
 756        "EXISTS": TokenType.EXISTS,
 757        "FALSE": TokenType.FALSE,
 758        "FETCH": TokenType.FETCH,
 759        "FILTER": TokenType.FILTER,
 760        "FIRST": TokenType.FIRST,
 761        "FULL": TokenType.FULL,
 762        "FUNCTION": TokenType.FUNCTION,
 763        "FOR": TokenType.FOR,
 764        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 765        "FORMAT": TokenType.FORMAT,
 766        "FROM": TokenType.FROM,
 767        "GEOGRAPHY": TokenType.GEOGRAPHY,
 768        "GEOMETRY": TokenType.GEOMETRY,
 769        "GLOB": TokenType.GLOB,
 770        "GROUP BY": TokenType.GROUP_BY,
 771        "GROUPING SETS": TokenType.GROUPING_SETS,
 772        "HAVING": TokenType.HAVING,
 773        "ILIKE": TokenType.ILIKE,
 774        "IN": TokenType.IN,
 775        "INDEX": TokenType.INDEX,
 776        "INET": TokenType.INET,
 777        "INNER": TokenType.INNER,
 778        "INSERT": TokenType.INSERT,
 779        "INTERVAL": TokenType.INTERVAL,
 780        "INTERSECT": TokenType.INTERSECT,
 781        "INTO": TokenType.INTO,
 782        "IS": TokenType.IS,
 783        "ISNULL": TokenType.ISNULL,
 784        "JOIN": TokenType.JOIN,
 785        "KEEP": TokenType.KEEP,
 786        "KILL": TokenType.KILL,
 787        "LATERAL": TokenType.LATERAL,
 788        "LEFT": TokenType.LEFT,
 789        "LIKE": TokenType.LIKE,
 790        "LIMIT": TokenType.LIMIT,
 791        "LOAD": TokenType.LOAD,
 792        "LOCK": TokenType.LOCK,
 793        "MERGE": TokenType.MERGE,
 794        "NAMESPACE": TokenType.NAMESPACE,
 795        "NATURAL": TokenType.NATURAL,
 796        "NEXT": TokenType.NEXT,
 797        "NOT": TokenType.NOT,
 798        "NOTNULL": TokenType.NOTNULL,
 799        "NULL": TokenType.NULL,
 800        "OBJECT": TokenType.OBJECT,
 801        "OFFSET": TokenType.OFFSET,
 802        "ON": TokenType.ON,
 803        "OR": TokenType.OR,
 804        "XOR": TokenType.XOR,
 805        "ORDER BY": TokenType.ORDER_BY,
 806        "ORDINALITY": TokenType.ORDINALITY,
 807        "OUTER": TokenType.OUTER,
 808        "OVER": TokenType.OVER,
 809        "OVERLAPS": TokenType.OVERLAPS,
 810        "OVERWRITE": TokenType.OVERWRITE,
 811        "PARTITION": TokenType.PARTITION,
 812        "PARTITION BY": TokenType.PARTITION_BY,
 813        "PARTITIONED BY": TokenType.PARTITION_BY,
 814        "PARTITIONED_BY": TokenType.PARTITION_BY,
 815        "PERCENT": TokenType.PERCENT,
 816        "PIVOT": TokenType.PIVOT,
 817        "PRAGMA": TokenType.PRAGMA,
 818        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 819        "PROCEDURE": TokenType.PROCEDURE,
 820        "QUALIFY": TokenType.QUALIFY,
 821        "RANGE": TokenType.RANGE,
 822        "RECURSIVE": TokenType.RECURSIVE,
 823        "REGEXP": TokenType.RLIKE,
 824        "RENAME": TokenType.RENAME,
 825        "REPLACE": TokenType.REPLACE,
 826        "RETURNING": TokenType.RETURNING,
 827        "REFERENCES": TokenType.REFERENCES,
 828        "RIGHT": TokenType.RIGHT,
 829        "RLIKE": TokenType.RLIKE,
 830        "ROLLBACK": TokenType.ROLLBACK,
 831        "ROLLUP": TokenType.ROLLUP,
 832        "ROW": TokenType.ROW,
 833        "ROWS": TokenType.ROWS,
 834        "SCHEMA": TokenType.SCHEMA,
 835        "SELECT": TokenType.SELECT,
 836        "SEMI": TokenType.SEMI,
 837        "SET": TokenType.SET,
 838        "SETTINGS": TokenType.SETTINGS,
 839        "SHOW": TokenType.SHOW,
 840        "SIMILAR TO": TokenType.SIMILAR_TO,
 841        "SOME": TokenType.SOME,
 842        "SORT BY": TokenType.SORT_BY,
 843        "START WITH": TokenType.START_WITH,
 844        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 845        "TABLE": TokenType.TABLE,
 846        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 847        "TEMP": TokenType.TEMPORARY,
 848        "TEMPORARY": TokenType.TEMPORARY,
 849        "THEN": TokenType.THEN,
 850        "TRUE": TokenType.TRUE,
 851        "TRUNCATE": TokenType.TRUNCATE,
 852        "UNION": TokenType.UNION,
 853        "UNKNOWN": TokenType.UNKNOWN,
 854        "UNNEST": TokenType.UNNEST,
 855        "UNPIVOT": TokenType.UNPIVOT,
 856        "UPDATE": TokenType.UPDATE,
 857        "USE": TokenType.USE,
 858        "USING": TokenType.USING,
 859        "UUID": TokenType.UUID,
 860        "VALUES": TokenType.VALUES,
 861        "VIEW": TokenType.VIEW,
 862        "VOLATILE": TokenType.VOLATILE,
 863        "WHEN": TokenType.WHEN,
 864        "WHERE": TokenType.WHERE,
 865        "WINDOW": TokenType.WINDOW,
 866        "WITH": TokenType.WITH,
 867        "APPLY": TokenType.APPLY,
 868        "ARRAY": TokenType.ARRAY,
 869        "BIT": TokenType.BIT,
 870        "BOOL": TokenType.BOOLEAN,
 871        "BOOLEAN": TokenType.BOOLEAN,
 872        "BYTE": TokenType.TINYINT,
 873        "MEDIUMINT": TokenType.MEDIUMINT,
 874        "INT1": TokenType.TINYINT,
 875        "TINYINT": TokenType.TINYINT,
 876        "INT16": TokenType.SMALLINT,
 877        "SHORT": TokenType.SMALLINT,
 878        "SMALLINT": TokenType.SMALLINT,
 879        "HUGEINT": TokenType.INT128,
 880        "UHUGEINT": TokenType.UINT128,
 881        "INT2": TokenType.SMALLINT,
 882        "INTEGER": TokenType.INT,
 883        "INT": TokenType.INT,
 884        "INT4": TokenType.INT,
 885        "INT32": TokenType.INT,
 886        "INT64": TokenType.BIGINT,
 887        "INT128": TokenType.INT128,
 888        "INT256": TokenType.INT256,
 889        "LONG": TokenType.BIGINT,
 890        "BIGINT": TokenType.BIGINT,
 891        "INT8": TokenType.TINYINT,
 892        "UINT": TokenType.UINT,
 893        "UINT128": TokenType.UINT128,
 894        "UINT256": TokenType.UINT256,
 895        "DEC": TokenType.DECIMAL,
 896        "DECIMAL": TokenType.DECIMAL,
 897        "DECIMAL32": TokenType.DECIMAL32,
 898        "DECIMAL64": TokenType.DECIMAL64,
 899        "DECIMAL128": TokenType.DECIMAL128,
 900        "DECIMAL256": TokenType.DECIMAL256,
 901        "BIGDECIMAL": TokenType.BIGDECIMAL,
 902        "BIGNUMERIC": TokenType.BIGDECIMAL,
 903        "LIST": TokenType.LIST,
 904        "MAP": TokenType.MAP,
 905        "NULLABLE": TokenType.NULLABLE,
 906        "NUMBER": TokenType.DECIMAL,
 907        "NUMERIC": TokenType.DECIMAL,
 908        "FIXED": TokenType.DECIMAL,
 909        "REAL": TokenType.FLOAT,
 910        "FLOAT": TokenType.FLOAT,
 911        "FLOAT4": TokenType.FLOAT,
 912        "FLOAT8": TokenType.DOUBLE,
 913        "DOUBLE": TokenType.DOUBLE,
 914        "DOUBLE PRECISION": TokenType.DOUBLE,
 915        "JSON": TokenType.JSON,
 916        "JSONB": TokenType.JSONB,
 917        "CHAR": TokenType.CHAR,
 918        "CHARACTER": TokenType.CHAR,
 919        "CHAR VARYING": TokenType.VARCHAR,
 920        "CHARACTER VARYING": TokenType.VARCHAR,
 921        "NCHAR": TokenType.NCHAR,
 922        "VARCHAR": TokenType.VARCHAR,
 923        "VARCHAR2": TokenType.VARCHAR,
 924        "NVARCHAR": TokenType.NVARCHAR,
 925        "NVARCHAR2": TokenType.NVARCHAR,
 926        "BPCHAR": TokenType.BPCHAR,
 927        "STR": TokenType.TEXT,
 928        "STRING": TokenType.TEXT,
 929        "TEXT": TokenType.TEXT,
 930        "LONGTEXT": TokenType.LONGTEXT,
 931        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 932        "TINYTEXT": TokenType.TINYTEXT,
 933        "CLOB": TokenType.TEXT,
 934        "LONGVARCHAR": TokenType.TEXT,
 935        "BINARY": TokenType.BINARY,
 936        "BLOB": TokenType.VARBINARY,
 937        "LONGBLOB": TokenType.LONGBLOB,
 938        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 939        "TINYBLOB": TokenType.TINYBLOB,
 940        "BYTEA": TokenType.VARBINARY,
 941        "VARBINARY": TokenType.VARBINARY,
 942        "TIME": TokenType.TIME,
 943        "TIMETZ": TokenType.TIMETZ,
 944        "TIMESTAMP": TokenType.TIMESTAMP,
 945        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 946        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 947        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 948        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 949        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 950        "DATE": TokenType.DATE,
 951        "DATETIME": TokenType.DATETIME,
 952        "INT4RANGE": TokenType.INT4RANGE,
 953        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 954        "INT8RANGE": TokenType.INT8RANGE,
 955        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 956        "NUMRANGE": TokenType.NUMRANGE,
 957        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 958        "TSRANGE": TokenType.TSRANGE,
 959        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 960        "TSTZRANGE": TokenType.TSTZRANGE,
 961        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 962        "DATERANGE": TokenType.DATERANGE,
 963        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 964        "UNIQUE": TokenType.UNIQUE,
 965        "VECTOR": TokenType.VECTOR,
 966        "STRUCT": TokenType.STRUCT,
 967        "SEQUENCE": TokenType.SEQUENCE,
 968        "VARIANT": TokenType.VARIANT,
 969        "ALTER": TokenType.ALTER,
 970        "ANALYZE": TokenType.ANALYZE,
 971        "CALL": TokenType.COMMAND,
 972        "COMMENT": TokenType.COMMENT,
 973        "EXPLAIN": TokenType.COMMAND,
 974        "GRANT": TokenType.GRANT,
 975        "OPTIMIZE": TokenType.COMMAND,
 976        "PREPARE": TokenType.COMMAND,
 977        "VACUUM": TokenType.COMMAND,
 978        "USER-DEFINED": TokenType.USERDEFINED,
 979        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 980        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 981    }
 982
 983    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 984        " ": TokenType.SPACE,
 985        "\t": TokenType.SPACE,
 986        "\n": TokenType.BREAK,
 987        "\r": TokenType.BREAK,
 988    }
 989
 990    COMMANDS = {
 991        TokenType.COMMAND,
 992        TokenType.EXECUTE,
 993        TokenType.FETCH,
 994        TokenType.SHOW,
 995        TokenType.RENAME,
 996    }
 997
 998    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 999
1000    # Handle numeric literals like in hive (3L = BIGINT)
1001    NUMERIC_LITERALS: t.Dict[str, str] = {}
1002
1003    COMMENTS = ["--", ("/*", "*/")]
1004
1005    __slots__ = (
1006        "sql",
1007        "size",
1008        "tokens",
1009        "dialect",
1010        "use_rs_tokenizer",
1011        "_start",
1012        "_current",
1013        "_line",
1014        "_col",
1015        "_comments",
1016        "_char",
1017        "_end",
1018        "_peek",
1019        "_prev_token_line",
1020        "_rs_dialect_settings",
1021    )
1022
1023    def __init__(
1024        self,
1025        dialect: DialectType = None,
1026        use_rs_tokenizer: t.Optional[bool] = None,
1027        **opts: t.Any,
1028    ) -> None:
1029        from sqlglot.dialects import Dialect
1030
1031        self.dialect = Dialect.get_or_raise(dialect)
1032
1033        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1034        self.use_rs_tokenizer = (
1035            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1036        )
1037
1038        if self.use_rs_tokenizer:
1039            self._rs_dialect_settings = RsTokenizerDialectSettings(
1040                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1041                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1042                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1043            )
1044
1045        self.reset()
1046
1047    def reset(self) -> None:
1048        self.sql = ""
1049        self.size = 0
1050        self.tokens: t.List[Token] = []
1051        self._start = 0
1052        self._current = 0
1053        self._line = 1
1054        self._col = 0
1055        self._comments: t.List[str] = []
1056
1057        self._char = ""
1058        self._end = False
1059        self._peek = ""
1060        self._prev_token_line = -1
1061
1062    def tokenize(self, sql: str) -> t.List[Token]:
1063        """Returns a list of tokens corresponding to the SQL string `sql`."""
1064        if self.use_rs_tokenizer:
1065            return self.tokenize_rs(sql)
1066
1067        self.reset()
1068        self.sql = sql
1069        self.size = len(sql)
1070
1071        try:
1072            self._scan()
1073        except Exception as e:
1074            start = max(self._current - 50, 0)
1075            end = min(self._current + 50, self.size - 1)
1076            context = self.sql[start:end]
1077            raise TokenError(f"Error tokenizing '{context}'") from e
1078
1079        return self.tokens
1080
1081    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1082        while self.size and not self._end:
1083            current = self._current
1084
1085            # Skip spaces here rather than iteratively calling advance() for performance reasons
1086            while current < self.size:
1087                char = self.sql[current]
1088
1089                if char.isspace() and (char == " " or char == "\t"):
1090                    current += 1
1091                else:
1092                    break
1093
1094            offset = current - self._current if current > self._current else 1
1095
1096            self._start = current
1097            self._advance(offset)
1098
1099            if not self._char.isspace():
1100                if self._char.isdigit():
1101                    self._scan_number()
1102                elif self._char in self._IDENTIFIERS:
1103                    self._scan_identifier(self._IDENTIFIERS[self._char])
1104                else:
1105                    self._scan_keywords()
1106
1107            if until and until():
1108                break
1109
1110        if self.tokens and self._comments:
1111            self.tokens[-1].comments.extend(self._comments)
1112
1113    def _chars(self, size: int) -> str:
1114        if size == 1:
1115            return self._char
1116
1117        start = self._current - 1
1118        end = start + size
1119
1120        return self.sql[start:end] if end <= self.size else ""
1121
1122    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1123        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1124            # Ensures we don't count an extra line if we get a \r\n line break sequence
1125            if not (self._char == "\r" and self._peek == "\n"):
1126                self._col = i
1127                self._line += 1
1128        else:
1129            self._col += i
1130
1131        self._current += i
1132        self._end = self._current >= self.size
1133        self._char = self.sql[self._current - 1]
1134        self._peek = "" if self._end else self.sql[self._current]
1135
1136        if alnum and self._char.isalnum():
1137            # Here we use local variables instead of attributes for better performance
1138            _col = self._col
1139            _current = self._current
1140            _end = self._end
1141            _peek = self._peek
1142
1143            while _peek.isalnum():
1144                _col += 1
1145                _current += 1
1146                _end = _current >= self.size
1147                _peek = "" if _end else self.sql[_current]
1148
1149            self._col = _col
1150            self._current = _current
1151            self._end = _end
1152            self._peek = _peek
1153            self._char = self.sql[_current - 1]
1154
1155    @property
1156    def _text(self) -> str:
1157        return self.sql[self._start : self._current]
1158
1159    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1160        self._prev_token_line = self._line
1161
1162        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1163            self.tokens[-1].comments.extend(self._comments)
1164            self._comments = []
1165
1166        self.tokens.append(
1167            Token(
1168                token_type,
1169                text=self._text if text is None else text,
1170                line=self._line,
1171                col=self._col,
1172                start=self._start,
1173                end=self._current - 1,
1174                comments=self._comments,
1175            )
1176        )
1177        self._comments = []
1178
1179        # If we have either a semicolon or a begin token before the command's token, we'll parse
1180        # whatever follows the command's token as a string
1181        if (
1182            token_type in self.COMMANDS
1183            and self._peek != ";"
1184            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1185        ):
1186            start = self._current
1187            tokens = len(self.tokens)
1188            self._scan(lambda: self._peek == ";")
1189            self.tokens = self.tokens[:tokens]
1190            text = self.sql[start : self._current].strip()
1191            if text:
1192                self._add(TokenType.STRING, text)
1193
1194    def _scan_keywords(self) -> None:
1195        size = 0
1196        word = None
1197        chars = self._text
1198        char = chars
1199        prev_space = False
1200        skip = False
1201        trie = self._KEYWORD_TRIE
1202        single_token = char in self.SINGLE_TOKENS
1203
1204        while chars:
1205            if skip:
1206                result = TrieResult.PREFIX
1207            else:
1208                result, trie = in_trie(trie, char.upper())
1209
1210            if result == TrieResult.FAILED:
1211                break
1212            if result == TrieResult.EXISTS:
1213                word = chars
1214
1215            end = self._current + size
1216            size += 1
1217
1218            if end < self.size:
1219                char = self.sql[end]
1220                single_token = single_token or char in self.SINGLE_TOKENS
1221                is_space = char.isspace()
1222
1223                if not is_space or not prev_space:
1224                    if is_space:
1225                        char = " "
1226                    chars += char
1227                    prev_space = is_space
1228                    skip = False
1229                else:
1230                    skip = True
1231            else:
1232                char = ""
1233                break
1234
1235        if word:
1236            if self._scan_string(word):
1237                return
1238            if self._scan_comment(word):
1239                return
1240            if prev_space or single_token or not char:
1241                self._advance(size - 1)
1242                word = word.upper()
1243                self._add(self.KEYWORDS[word], text=word)
1244                return
1245
1246        if self._char in self.SINGLE_TOKENS:
1247            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1248            return
1249
1250        self._scan_var()
1251
1252    def _scan_comment(self, comment_start: str) -> bool:
1253        if comment_start not in self._COMMENTS:
1254            return False
1255
1256        comment_start_line = self._line
1257        comment_start_size = len(comment_start)
1258        comment_end = self._COMMENTS[comment_start]
1259
1260        if comment_end:
1261            # Skip the comment's start delimiter
1262            self._advance(comment_start_size)
1263
1264            comment_count = 1
1265            comment_end_size = len(comment_end)
1266
1267            while not self._end:
1268                if self._chars(comment_end_size) == comment_end:
1269                    comment_count -= 1
1270                    if not comment_count:
1271                        break
1272
1273                self._advance(alnum=True)
1274
1275                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1276                if (
1277                    self.NESTED_COMMENTS
1278                    and not self._end
1279                    and self._chars(comment_end_size) == comment_start
1280                ):
1281                    self._advance(comment_start_size)
1282                    comment_count += 1
1283
1284            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1285            self._advance(comment_end_size - 1)
1286        else:
1287            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1288                self._advance(alnum=True)
1289            self._comments.append(self._text[comment_start_size:])
1290
1291        if (
1292            comment_start == self.HINT_START
1293            and self.tokens
1294            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1295        ):
1296            self._add(TokenType.HINT)
1297
1298        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1299        # Multiple consecutive comments are preserved by appending them to the current comments list.
1300        if comment_start_line == self._prev_token_line:
1301            self.tokens[-1].comments.extend(self._comments)
1302            self._comments = []
1303            self._prev_token_line = self._line
1304
1305        return True
1306
1307    def _scan_number(self) -> None:
1308        if self._char == "0":
1309            peek = self._peek.upper()
1310            if peek == "B":
1311                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1312            elif peek == "X":
1313                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1314
1315        decimal = False
1316        scientific = 0
1317
1318        while True:
1319            if self._peek.isdigit():
1320                self._advance()
1321            elif self._peek == "." and not decimal:
1322                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1323                    return self._add(TokenType.NUMBER)
1324                decimal = True
1325                self._advance()
1326            elif self._peek in ("-", "+") and scientific == 1:
1327                scientific += 1
1328                self._advance()
1329            elif self._peek.upper() == "E" and not scientific:
1330                scientific += 1
1331                self._advance()
1332            elif self._peek.isidentifier():
1333                number_text = self._text
1334                literal = ""
1335
1336                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1337                    literal += self._peek
1338                    self._advance()
1339
1340                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1341
1342                if token_type:
1343                    self._add(TokenType.NUMBER, number_text)
1344                    self._add(TokenType.DCOLON, "::")
1345                    return self._add(token_type, literal)
1346                else:
1347                    replaced = literal.replace("_", "")
1348                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1349                        return self._add(TokenType.NUMBER, number_text + replaced)
1350                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1351                        return self._add(TokenType.VAR)
1352
1353                self._advance(-len(literal))
1354                return self._add(TokenType.NUMBER, number_text)
1355            else:
1356                return self._add(TokenType.NUMBER)
1357
1358    def _scan_bits(self) -> None:
1359        self._advance()
1360        value = self._extract_value()
1361        try:
1362            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1363            int(value, 2)
1364            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1365        except ValueError:
1366            self._add(TokenType.IDENTIFIER)
1367
1368    def _scan_hex(self) -> None:
1369        self._advance()
1370        value = self._extract_value()
1371        try:
1372            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1373            int(value, 16)
1374            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1375        except ValueError:
1376            self._add(TokenType.IDENTIFIER)
1377
1378    def _extract_value(self) -> str:
1379        while True:
1380            char = self._peek.strip()
1381            if char and char not in self.SINGLE_TOKENS:
1382                self._advance(alnum=True)
1383            else:
1384                break
1385
1386        return self._text
1387
1388    def _scan_string(self, start: str) -> bool:
1389        base = None
1390        token_type = TokenType.STRING
1391
1392        if start in self._QUOTES:
1393            end = self._QUOTES[start]
1394        elif start in self._FORMAT_STRINGS:
1395            end, token_type = self._FORMAT_STRINGS[start]
1396
1397            if token_type == TokenType.HEX_STRING:
1398                base = 16
1399            elif token_type == TokenType.BIT_STRING:
1400                base = 2
1401            elif token_type == TokenType.HEREDOC_STRING:
1402                self._advance()
1403
1404                if self._char == end:
1405                    tag = ""
1406                else:
1407                    tag = self._extract_string(
1408                        end,
1409                        raw_string=True,
1410                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1411                    )
1412
1413                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1414                    if not self._end:
1415                        self._advance(-1)
1416
1417                    self._advance(-len(tag))
1418                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1419                    return True
1420
1421                end = f"{start}{tag}{end}"
1422        else:
1423            return False
1424
1425        self._advance(len(start))
1426        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1427
1428        if base:
1429            try:
1430                int(text, base)
1431            except Exception:
1432                raise TokenError(
1433                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1434                )
1435
1436        self._add(token_type, text)
1437        return True
1438
1439    def _scan_identifier(self, identifier_end: str) -> None:
1440        self._advance()
1441        text = self._extract_string(
1442            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1443        )
1444        self._add(TokenType.IDENTIFIER, text)
1445
1446    def _scan_var(self) -> None:
1447        while True:
1448            char = self._peek.strip()
1449            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1450                self._advance(alnum=True)
1451            else:
1452                break
1453
1454        self._add(
1455            TokenType.VAR
1456            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1457            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1458        )
1459
1460    def _extract_string(
1461        self,
1462        delimiter: str,
1463        escapes: t.Optional[t.Set[str]] = None,
1464        raw_string: bool = False,
1465        raise_unmatched: bool = True,
1466    ) -> str:
1467        text = ""
1468        delim_size = len(delimiter)
1469        escapes = self._STRING_ESCAPES if escapes is None else escapes
1470
1471        while True:
1472            if (
1473                not raw_string
1474                and self.dialect.UNESCAPED_SEQUENCES
1475                and self._peek
1476                and self._char in self.STRING_ESCAPES
1477            ):
1478                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1479                if unescaped_sequence:
1480                    self._advance(2)
1481                    text += unescaped_sequence
1482                    continue
1483            if (
1484                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1485                and self._char in escapes
1486                and (self._peek == delimiter or self._peek in escapes)
1487                and (self._char not in self._QUOTES or self._char == self._peek)
1488            ):
1489                if self._peek == delimiter:
1490                    text += self._peek
1491                else:
1492                    text += self._char + self._peek
1493
1494                if self._current + 1 < self.size:
1495                    self._advance(2)
1496                else:
1497                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1498            else:
1499                if self._chars(delim_size) == delimiter:
1500                    if delim_size > 1:
1501                        self._advance(delim_size - 1)
1502                    break
1503
1504                if self._end:
1505                    if not raise_unmatched:
1506                        return text + self._char
1507
1508                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1509
1510                current = self._current - 1
1511                self._advance(alnum=True)
1512                text += self.sql[current : self._current - 1]
1513
1514        return text
1515
1516    def tokenize_rs(self, sql: str) -> t.List[Token]:
1517        if not self._RS_TOKENIZER:
1518            raise SqlglotError("Rust tokenizer is not available")
1519
1520        tokens, error_msg = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1521        for token in tokens:
1522            token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1523
1524        # Setting this here so partial token lists can be inspected even if there is a failure
1525        self.tokens = tokens
1526
1527        if error_msg is not None:
1528            raise TokenError(error_msg)
1529
1530        return tokens