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