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