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