sqlglot.tokens

   1from __future__ import annotations
   2
   3import os
   4import typing as t
   5from enum import auto
   6
   7from sqlglot.errors import SqlglotError, TokenError
   8from sqlglot.helper import AutoName
   9from sqlglot.trie import TrieResult, in_trie, new_trie
  10
  11if t.TYPE_CHECKING:
  12    from sqlglot.dialects.dialect import DialectType
  13
  14
  15try:
  16    from sqlglotrs import (  # type: ignore
  17        Tokenizer as RsTokenizer,
  18        TokenizerDialectSettings as RsTokenizerDialectSettings,
  19        TokenizerSettings as RsTokenizerSettings,
  20        TokenTypeSettings as RsTokenTypeSettings,
  21    )
  22
  23    USE_RS_TOKENIZER = os.environ.get("SQLGLOTRS_TOKENIZER", "1") == "1"
  24except ImportError:
  25    USE_RS_TOKENIZER = False
  26
  27
  28class TokenType(AutoName):
  29    L_PAREN = auto()
  30    R_PAREN = auto()
  31    L_BRACKET = auto()
  32    R_BRACKET = auto()
  33    L_BRACE = auto()
  34    R_BRACE = auto()
  35    COMMA = auto()
  36    DOT = auto()
  37    DASH = auto()
  38    PLUS = auto()
  39    COLON = auto()
  40    DOTCOLON = auto()
  41    DCOLON = auto()
  42    DQMARK = auto()
  43    SEMICOLON = auto()
  44    STAR = auto()
  45    BACKSLASH = auto()
  46    SLASH = auto()
  47    LT = auto()
  48    LTE = auto()
  49    GT = auto()
  50    GTE = auto()
  51    NOT = auto()
  52    EQ = auto()
  53    NEQ = auto()
  54    NULLSAFE_EQ = auto()
  55    COLON_EQ = auto()
  56    AND = auto()
  57    OR = auto()
  58    AMP = auto()
  59    DPIPE = auto()
  60    PIPE = auto()
  61    PIPE_SLASH = auto()
  62    DPIPE_SLASH = auto()
  63    CARET = auto()
  64    CARET_AT = auto()
  65    TILDA = auto()
  66    ARROW = auto()
  67    DARROW = auto()
  68    FARROW = auto()
  69    HASH = auto()
  70    HASH_ARROW = auto()
  71    DHASH_ARROW = auto()
  72    LR_ARROW = auto()
  73    DAT = auto()
  74    LT_AT = auto()
  75    AT_GT = auto()
  76    DOLLAR = auto()
  77    PARAMETER = auto()
  78    SESSION_PARAMETER = auto()
  79    DAMP = auto()
  80    XOR = auto()
  81    DSTAR = auto()
  82
  83    URI_START = auto()
  84
  85    BLOCK_START = auto()
  86    BLOCK_END = auto()
  87
  88    SPACE = auto()
  89    BREAK = auto()
  90
  91    STRING = auto()
  92    NUMBER = auto()
  93    IDENTIFIER = auto()
  94    DATABASE = auto()
  95    COLUMN = auto()
  96    COLUMN_DEF = auto()
  97    SCHEMA = auto()
  98    TABLE = auto()
  99    WAREHOUSE = auto()
 100    STREAMLIT = auto()
 101    VAR = auto()
 102    BIT_STRING = auto()
 103    HEX_STRING = auto()
 104    BYTE_STRING = auto()
 105    NATIONAL_STRING = auto()
 106    RAW_STRING = auto()
 107    HEREDOC_STRING = auto()
 108    UNICODE_STRING = auto()
 109
 110    # types
 111    BIT = auto()
 112    BOOLEAN = auto()
 113    TINYINT = auto()
 114    UTINYINT = auto()
 115    SMALLINT = auto()
 116    USMALLINT = auto()
 117    MEDIUMINT = auto()
 118    UMEDIUMINT = auto()
 119    INT = auto()
 120    UINT = auto()
 121    BIGINT = auto()
 122    UBIGINT = auto()
 123    INT128 = auto()
 124    UINT128 = auto()
 125    INT256 = auto()
 126    UINT256 = auto()
 127    FLOAT = auto()
 128    DOUBLE = auto()
 129    UDOUBLE = auto()
 130    DECIMAL = auto()
 131    DECIMAL32 = auto()
 132    DECIMAL64 = auto()
 133    DECIMAL128 = auto()
 134    DECIMAL256 = auto()
 135    UDECIMAL = auto()
 136    BIGDECIMAL = auto()
 137    CHAR = auto()
 138    NCHAR = auto()
 139    VARCHAR = auto()
 140    NVARCHAR = auto()
 141    BPCHAR = auto()
 142    TEXT = auto()
 143    MEDIUMTEXT = auto()
 144    LONGTEXT = auto()
 145    BLOB = auto()
 146    MEDIUMBLOB = auto()
 147    LONGBLOB = auto()
 148    TINYBLOB = auto()
 149    TINYTEXT = auto()
 150    NAME = auto()
 151    BINARY = auto()
 152    VARBINARY = auto()
 153    JSON = auto()
 154    JSONB = auto()
 155    TIME = auto()
 156    TIMETZ = auto()
 157    TIMESTAMP = auto()
 158    TIMESTAMPTZ = auto()
 159    TIMESTAMPLTZ = auto()
 160    TIMESTAMPNTZ = auto()
 161    TIMESTAMP_S = auto()
 162    TIMESTAMP_MS = auto()
 163    TIMESTAMP_NS = auto()
 164    DATETIME = auto()
 165    DATETIME2 = auto()
 166    DATETIME64 = auto()
 167    SMALLDATETIME = auto()
 168    DATE = auto()
 169    DATE32 = auto()
 170    INT4RANGE = auto()
 171    INT4MULTIRANGE = auto()
 172    INT8RANGE = auto()
 173    INT8MULTIRANGE = auto()
 174    NUMRANGE = auto()
 175    NUMMULTIRANGE = auto()
 176    TSRANGE = auto()
 177    TSMULTIRANGE = auto()
 178    TSTZRANGE = auto()
 179    TSTZMULTIRANGE = auto()
 180    DATERANGE = auto()
 181    DATEMULTIRANGE = auto()
 182    UUID = auto()
 183    GEOGRAPHY = auto()
 184    NULLABLE = auto()
 185    GEOMETRY = auto()
 186    POINT = auto()
 187    RING = auto()
 188    LINESTRING = auto()
 189    MULTILINESTRING = auto()
 190    POLYGON = auto()
 191    MULTIPOLYGON = auto()
 192    HLLSKETCH = auto()
 193    HSTORE = auto()
 194    SUPER = auto()
 195    SERIAL = auto()
 196    SMALLSERIAL = auto()
 197    BIGSERIAL = auto()
 198    XML = auto()
 199    YEAR = auto()
 200    USERDEFINED = auto()
 201    MONEY = auto()
 202    SMALLMONEY = auto()
 203    ROWVERSION = auto()
 204    IMAGE = auto()
 205    VARIANT = auto()
 206    OBJECT = auto()
 207    INET = auto()
 208    IPADDRESS = auto()
 209    IPPREFIX = auto()
 210    IPV4 = auto()
 211    IPV6 = auto()
 212    ENUM = auto()
 213    ENUM8 = auto()
 214    ENUM16 = auto()
 215    FIXEDSTRING = auto()
 216    LOWCARDINALITY = auto()
 217    NESTED = auto()
 218    AGGREGATEFUNCTION = auto()
 219    SIMPLEAGGREGATEFUNCTION = auto()
 220    TDIGEST = auto()
 221    UNKNOWN = auto()
 222    VECTOR = auto()
 223    DYNAMIC = auto()
 224
 225    # keywords
 226    ALIAS = auto()
 227    ALTER = auto()
 228    ALWAYS = auto()
 229    ALL = auto()
 230    ANTI = auto()
 231    ANY = auto()
 232    APPLY = auto()
 233    ARRAY = auto()
 234    ASC = auto()
 235    ASOF = auto()
 236    ATTACH = auto()
 237    AUTO_INCREMENT = auto()
 238    BEGIN = auto()
 239    BETWEEN = auto()
 240    BULK_COLLECT_INTO = auto()
 241    CACHE = auto()
 242    CASE = auto()
 243    CHARACTER_SET = auto()
 244    CLUSTER_BY = auto()
 245    COLLATE = auto()
 246    COMMAND = auto()
 247    COMMENT = auto()
 248    COMMIT = auto()
 249    CONNECT_BY = auto()
 250    CONSTRAINT = auto()
 251    COPY = auto()
 252    CREATE = auto()
 253    CROSS = auto()
 254    CUBE = auto()
 255    CURRENT_DATE = auto()
 256    CURRENT_DATETIME = auto()
 257    CURRENT_SCHEMA = auto()
 258    CURRENT_TIME = auto()
 259    CURRENT_TIMESTAMP = auto()
 260    CURRENT_USER = auto()
 261    DECLARE = auto()
 262    DEFAULT = auto()
 263    DELETE = auto()
 264    DESC = auto()
 265    DESCRIBE = auto()
 266    DETACH = auto()
 267    DICTIONARY = auto()
 268    DISTINCT = auto()
 269    DISTRIBUTE_BY = auto()
 270    DIV = auto()
 271    DROP = auto()
 272    ELSE = auto()
 273    END = auto()
 274    ESCAPE = auto()
 275    EXCEPT = auto()
 276    EXECUTE = auto()
 277    EXISTS = auto()
 278    FALSE = auto()
 279    FETCH = auto()
 280    FILTER = auto()
 281    FINAL = auto()
 282    FIRST = auto()
 283    FOR = auto()
 284    FORCE = auto()
 285    FOREIGN_KEY = auto()
 286    FORMAT = auto()
 287    FROM = auto()
 288    FULL = auto()
 289    FUNCTION = auto()
 290    GLOB = auto()
 291    GLOBAL = auto()
 292    GRANT = auto()
 293    GROUP_BY = auto()
 294    GROUPING_SETS = auto()
 295    HAVING = auto()
 296    HINT = auto()
 297    IGNORE = auto()
 298    ILIKE = auto()
 299    ILIKE_ANY = auto()
 300    IN = auto()
 301    INDEX = auto()
 302    INNER = auto()
 303    INSERT = auto()
 304    INTERSECT = auto()
 305    INTERVAL = auto()
 306    INTO = auto()
 307    INTRODUCER = auto()
 308    IRLIKE = auto()
 309    IS = auto()
 310    ISNULL = auto()
 311    JOIN = auto()
 312    JOIN_MARKER = auto()
 313    KEEP = auto()
 314    KEY = auto()
 315    KILL = auto()
 316    LANGUAGE = auto()
 317    LATERAL = auto()
 318    LEFT = auto()
 319    LIKE = auto()
 320    LIKE_ANY = auto()
 321    LIMIT = auto()
 322    LIST = auto()
 323    LOAD = auto()
 324    LOCK = auto()
 325    MAP = auto()
 326    MATCH_CONDITION = auto()
 327    MATCH_RECOGNIZE = auto()
 328    MEMBER_OF = auto()
 329    MERGE = auto()
 330    MOD = auto()
 331    MODEL = auto()
 332    NATURAL = auto()
 333    NEXT = auto()
 334    NOTNULL = auto()
 335    NULL = auto()
 336    OBJECT_IDENTIFIER = auto()
 337    OFFSET = auto()
 338    ON = auto()
 339    ONLY = auto()
 340    OPERATOR = auto()
 341    ORDER_BY = auto()
 342    ORDER_SIBLINGS_BY = auto()
 343    ORDERED = auto()
 344    ORDINALITY = auto()
 345    OUTER = auto()
 346    OVER = auto()
 347    OVERLAPS = auto()
 348    OVERWRITE = auto()
 349    PARTITION = auto()
 350    PARTITION_BY = auto()
 351    PERCENT = auto()
 352    PIVOT = auto()
 353    PLACEHOLDER = auto()
 354    POSITIONAL = auto()
 355    PRAGMA = auto()
 356    PREWHERE = auto()
 357    PRIMARY_KEY = auto()
 358    PROCEDURE = auto()
 359    PROPERTIES = auto()
 360    PSEUDO_TYPE = auto()
 361    PUT = auto()
 362    QUALIFY = auto()
 363    QUOTE = auto()
 364    RANGE = auto()
 365    RECURSIVE = auto()
 366    REFRESH = auto()
 367    RENAME = auto()
 368    REPLACE = auto()
 369    RETURNING = auto()
 370    REFERENCES = auto()
 371    RIGHT = auto()
 372    RLIKE = auto()
 373    ROLLBACK = auto()
 374    ROLLUP = auto()
 375    ROW = auto()
 376    ROWS = auto()
 377    SELECT = auto()
 378    SEMI = auto()
 379    SEPARATOR = auto()
 380    SEQUENCE = auto()
 381    SERDE_PROPERTIES = auto()
 382    SET = auto()
 383    SETTINGS = auto()
 384    SHOW = auto()
 385    SIMILAR_TO = auto()
 386    SOME = auto()
 387    SORT_BY = auto()
 388    START_WITH = auto()
 389    STORAGE_INTEGRATION = auto()
 390    STRAIGHT_JOIN = auto()
 391    STRUCT = auto()
 392    SUMMARIZE = auto()
 393    TABLE_SAMPLE = auto()
 394    TAG = auto()
 395    TEMPORARY = auto()
 396    TOP = auto()
 397    THEN = auto()
 398    TRUE = auto()
 399    TRUNCATE = auto()
 400    UNCACHE = auto()
 401    UNION = auto()
 402    UNNEST = auto()
 403    UNPIVOT = auto()
 404    UPDATE = auto()
 405    USE = auto()
 406    USING = auto()
 407    VALUES = auto()
 408    VIEW = auto()
 409    VOLATILE = auto()
 410    WHEN = auto()
 411    WHERE = auto()
 412    WINDOW = auto()
 413    WITH = auto()
 414    UNIQUE = auto()
 415    VERSION_SNAPSHOT = auto()
 416    TIMESTAMP_SNAPSHOT = auto()
 417    OPTION = auto()
 418    SINK = auto()
 419    SOURCE = auto()
 420    ANALYZE = auto()
 421    NAMESPACE = auto()
 422    EXPORT = auto()
 423
 424
 425_ALL_TOKEN_TYPES = list(TokenType)
 426_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 427
 428
 429class Token:
 430    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 431
 432    @classmethod
 433    def number(cls, number: int) -> Token:
 434        """Returns a NUMBER token with `number` as its text."""
 435        return cls(TokenType.NUMBER, str(number))
 436
 437    @classmethod
 438    def string(cls, string: str) -> Token:
 439        """Returns a STRING token with `string` as its text."""
 440        return cls(TokenType.STRING, string)
 441
 442    @classmethod
 443    def identifier(cls, identifier: str) -> Token:
 444        """Returns an IDENTIFIER token with `identifier` as its text."""
 445        return cls(TokenType.IDENTIFIER, identifier)
 446
 447    @classmethod
 448    def var(cls, var: str) -> Token:
 449        """Returns an VAR token with `var` as its text."""
 450        return cls(TokenType.VAR, var)
 451
 452    def __init__(
 453        self,
 454        token_type: TokenType,
 455        text: str,
 456        line: int = 1,
 457        col: int = 1,
 458        start: int = 0,
 459        end: int = 0,
 460        comments: t.Optional[t.List[str]] = None,
 461    ) -> None:
 462        """Token initializer.
 463
 464        Args:
 465            token_type: The TokenType Enum.
 466            text: The text of the token.
 467            line: The line that the token ends on.
 468            col: The column that the token ends on.
 469            start: The start index of the token.
 470            end: The ending index of the token.
 471            comments: The comments to attach to the token.
 472        """
 473        self.token_type = token_type
 474        self.text = text
 475        self.line = line
 476        self.col = col
 477        self.start = start
 478        self.end = end
 479        self.comments = [] if comments is None else comments
 480
 481    def __repr__(self) -> str:
 482        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 483        return f"<Token {attributes}>"
 484
 485
 486class _Tokenizer(type):
 487    def __new__(cls, clsname, bases, attrs):
 488        klass = super().__new__(cls, clsname, bases, attrs)
 489
 490        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 491            return dict(
 492                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 493            )
 494
 495        def _quotes_to_format(
 496            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 497        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 498            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 499
 500        klass._QUOTES = _convert_quotes(klass.QUOTES)
 501        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 502
 503        klass._FORMAT_STRINGS = {
 504            **{
 505                p + s: (e, TokenType.NATIONAL_STRING)
 506                for s, e in klass._QUOTES.items()
 507                for p in ("n", "N")
 508            },
 509            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 510            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 511            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 512            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 513            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 514            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 515        }
 516
 517        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 518        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 519        klass._COMMENTS = {
 520            **dict(
 521                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 522                for comment in klass.COMMENTS
 523            ),
 524            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 525        }
 526        if klass.HINT_START in klass.KEYWORDS:
 527            klass._COMMENTS[klass.HINT_START] = "*/"
 528
 529        klass._KEYWORD_TRIE = new_trie(
 530            key.upper()
 531            for key in (
 532                *klass.KEYWORDS,
 533                *klass._COMMENTS,
 534                *klass._QUOTES,
 535                *klass._FORMAT_STRINGS,
 536            )
 537            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 538        )
 539
 540        if USE_RS_TOKENIZER:
 541            settings = RsTokenizerSettings(
 542                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 543                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 544                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 545                numeric_literals=klass.NUMERIC_LITERALS,
 546                identifiers=klass._IDENTIFIERS,
 547                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 548                string_escapes=klass._STRING_ESCAPES,
 549                quotes=klass._QUOTES,
 550                format_strings={
 551                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 552                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 553                },
 554                has_bit_strings=bool(klass.BIT_STRINGS),
 555                has_hex_strings=bool(klass.HEX_STRINGS),
 556                comments=klass._COMMENTS,
 557                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 558                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 559                command_prefix_tokens={
 560                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 561                },
 562                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 563                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 564                nested_comments=klass.NESTED_COMMENTS,
 565                hint_start=klass.HINT_START,
 566                tokens_preceding_hint={
 567                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 568                },
 569            )
 570            token_types = RsTokenTypeSettings(
 571                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 572                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 573                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 574                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 575                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 576                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 577                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 578                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 579                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 580                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 581                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 582                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 583                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 584                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 585            )
 586            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 587        else:
 588            klass._RS_TOKENIZER = None
 589
 590        return klass
 591
 592
 593class Tokenizer(metaclass=_Tokenizer):
 594    SINGLE_TOKENS = {
 595        "(": TokenType.L_PAREN,
 596        ")": TokenType.R_PAREN,
 597        "[": TokenType.L_BRACKET,
 598        "]": TokenType.R_BRACKET,
 599        "{": TokenType.L_BRACE,
 600        "}": TokenType.R_BRACE,
 601        "&": TokenType.AMP,
 602        "^": TokenType.CARET,
 603        ":": TokenType.COLON,
 604        ",": TokenType.COMMA,
 605        ".": TokenType.DOT,
 606        "-": TokenType.DASH,
 607        "=": TokenType.EQ,
 608        ">": TokenType.GT,
 609        "<": TokenType.LT,
 610        "%": TokenType.MOD,
 611        "!": TokenType.NOT,
 612        "|": TokenType.PIPE,
 613        "+": TokenType.PLUS,
 614        ";": TokenType.SEMICOLON,
 615        "/": TokenType.SLASH,
 616        "\\": TokenType.BACKSLASH,
 617        "*": TokenType.STAR,
 618        "~": TokenType.TILDA,
 619        "?": TokenType.PLACEHOLDER,
 620        "@": TokenType.PARAMETER,
 621        "#": TokenType.HASH,
 622        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 623        "'": TokenType.UNKNOWN,
 624        "`": TokenType.UNKNOWN,
 625        '"': TokenType.UNKNOWN,
 626    }
 627
 628    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 629    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 630    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 631    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 632    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 633    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 634    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 635    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 636    STRING_ESCAPES = ["'"]
 637    VAR_SINGLE_TOKENS: t.Set[str] = set()
 638
 639    # The strings in this list can always be used as escapes, regardless of the surrounding
 640    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 641    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 642    IDENTIFIER_ESCAPES: t.List[str] = []
 643
 644    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 645    HEREDOC_TAG_IS_IDENTIFIER = False
 646
 647    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 648    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 649
 650    # Whether string escape characters function as such when placed within raw strings
 651    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 652
 653    NESTED_COMMENTS = True
 654
 655    HINT_START = "/*+"
 656
 657    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 658
 659    # Autofilled
 660    _COMMENTS: t.Dict[str, str] = {}
 661    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 662    _IDENTIFIERS: t.Dict[str, str] = {}
 663    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 664    _QUOTES: t.Dict[str, str] = {}
 665    _STRING_ESCAPES: t.Set[str] = set()
 666    _KEYWORD_TRIE: t.Dict = {}
 667    _RS_TOKENIZER: t.Optional[t.Any] = None
 668
 669    KEYWORDS: t.Dict[str, TokenType] = {
 670        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 671        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 672        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 673        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 674        HINT_START: TokenType.HINT,
 675        "==": TokenType.EQ,
 676        "::": TokenType.DCOLON,
 677        "||": TokenType.DPIPE,
 678        ">=": TokenType.GTE,
 679        "<=": TokenType.LTE,
 680        "<>": TokenType.NEQ,
 681        "!=": TokenType.NEQ,
 682        ":=": TokenType.COLON_EQ,
 683        "<=>": TokenType.NULLSAFE_EQ,
 684        "->": TokenType.ARROW,
 685        "->>": TokenType.DARROW,
 686        "=>": TokenType.FARROW,
 687        "#>": TokenType.HASH_ARROW,
 688        "#>>": TokenType.DHASH_ARROW,
 689        "<->": TokenType.LR_ARROW,
 690        "&&": TokenType.DAMP,
 691        "??": TokenType.DQMARK,
 692        "~~~": TokenType.GLOB,
 693        "~~": TokenType.LIKE,
 694        "~~*": TokenType.ILIKE,
 695        "~*": TokenType.IRLIKE,
 696        "ALL": TokenType.ALL,
 697        "ALWAYS": TokenType.ALWAYS,
 698        "AND": TokenType.AND,
 699        "ANTI": TokenType.ANTI,
 700        "ANY": TokenType.ANY,
 701        "ASC": TokenType.ASC,
 702        "AS": TokenType.ALIAS,
 703        "ASOF": TokenType.ASOF,
 704        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 705        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 706        "BEGIN": TokenType.BEGIN,
 707        "BETWEEN": TokenType.BETWEEN,
 708        "CACHE": TokenType.CACHE,
 709        "UNCACHE": TokenType.UNCACHE,
 710        "CASE": TokenType.CASE,
 711        "CHARACTER SET": TokenType.CHARACTER_SET,
 712        "CLUSTER BY": TokenType.CLUSTER_BY,
 713        "COLLATE": TokenType.COLLATE,
 714        "COLUMN": TokenType.COLUMN,
 715        "COMMIT": TokenType.COMMIT,
 716        "CONNECT BY": TokenType.CONNECT_BY,
 717        "CONSTRAINT": TokenType.CONSTRAINT,
 718        "COPY": TokenType.COPY,
 719        "CREATE": TokenType.CREATE,
 720        "CROSS": TokenType.CROSS,
 721        "CUBE": TokenType.CUBE,
 722        "CURRENT_DATE": TokenType.CURRENT_DATE,
 723        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 724        "CURRENT_TIME": TokenType.CURRENT_TIME,
 725        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 726        "CURRENT_USER": TokenType.CURRENT_USER,
 727        "DATABASE": TokenType.DATABASE,
 728        "DEFAULT": TokenType.DEFAULT,
 729        "DELETE": TokenType.DELETE,
 730        "DESC": TokenType.DESC,
 731        "DESCRIBE": TokenType.DESCRIBE,
 732        "DISTINCT": TokenType.DISTINCT,
 733        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 734        "DIV": TokenType.DIV,
 735        "DROP": TokenType.DROP,
 736        "ELSE": TokenType.ELSE,
 737        "END": TokenType.END,
 738        "ENUM": TokenType.ENUM,
 739        "ESCAPE": TokenType.ESCAPE,
 740        "EXCEPT": TokenType.EXCEPT,
 741        "EXECUTE": TokenType.EXECUTE,
 742        "EXISTS": TokenType.EXISTS,
 743        "FALSE": TokenType.FALSE,
 744        "FETCH": TokenType.FETCH,
 745        "FILTER": TokenType.FILTER,
 746        "FIRST": TokenType.FIRST,
 747        "FULL": TokenType.FULL,
 748        "FUNCTION": TokenType.FUNCTION,
 749        "FOR": TokenType.FOR,
 750        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 751        "FORMAT": TokenType.FORMAT,
 752        "FROM": TokenType.FROM,
 753        "GEOGRAPHY": TokenType.GEOGRAPHY,
 754        "GEOMETRY": TokenType.GEOMETRY,
 755        "GLOB": TokenType.GLOB,
 756        "GROUP BY": TokenType.GROUP_BY,
 757        "GROUPING SETS": TokenType.GROUPING_SETS,
 758        "HAVING": TokenType.HAVING,
 759        "ILIKE": TokenType.ILIKE,
 760        "IN": TokenType.IN,
 761        "INDEX": TokenType.INDEX,
 762        "INET": TokenType.INET,
 763        "INNER": TokenType.INNER,
 764        "INSERT": TokenType.INSERT,
 765        "INTERVAL": TokenType.INTERVAL,
 766        "INTERSECT": TokenType.INTERSECT,
 767        "INTO": TokenType.INTO,
 768        "IS": TokenType.IS,
 769        "ISNULL": TokenType.ISNULL,
 770        "JOIN": TokenType.JOIN,
 771        "KEEP": TokenType.KEEP,
 772        "KILL": TokenType.KILL,
 773        "LATERAL": TokenType.LATERAL,
 774        "LEFT": TokenType.LEFT,
 775        "LIKE": TokenType.LIKE,
 776        "LIMIT": TokenType.LIMIT,
 777        "LOAD": TokenType.LOAD,
 778        "LOCK": TokenType.LOCK,
 779        "MERGE": TokenType.MERGE,
 780        "NAMESPACE": TokenType.NAMESPACE,
 781        "NATURAL": TokenType.NATURAL,
 782        "NEXT": TokenType.NEXT,
 783        "NOT": TokenType.NOT,
 784        "NOTNULL": TokenType.NOTNULL,
 785        "NULL": TokenType.NULL,
 786        "OBJECT": TokenType.OBJECT,
 787        "OFFSET": TokenType.OFFSET,
 788        "ON": TokenType.ON,
 789        "OR": TokenType.OR,
 790        "XOR": TokenType.XOR,
 791        "ORDER BY": TokenType.ORDER_BY,
 792        "ORDINALITY": TokenType.ORDINALITY,
 793        "OUTER": TokenType.OUTER,
 794        "OVER": TokenType.OVER,
 795        "OVERLAPS": TokenType.OVERLAPS,
 796        "OVERWRITE": TokenType.OVERWRITE,
 797        "PARTITION": TokenType.PARTITION,
 798        "PARTITION BY": TokenType.PARTITION_BY,
 799        "PARTITIONED BY": TokenType.PARTITION_BY,
 800        "PARTITIONED_BY": TokenType.PARTITION_BY,
 801        "PERCENT": TokenType.PERCENT,
 802        "PIVOT": TokenType.PIVOT,
 803        "PRAGMA": TokenType.PRAGMA,
 804        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 805        "PROCEDURE": TokenType.PROCEDURE,
 806        "QUALIFY": TokenType.QUALIFY,
 807        "RANGE": TokenType.RANGE,
 808        "RECURSIVE": TokenType.RECURSIVE,
 809        "REGEXP": TokenType.RLIKE,
 810        "RENAME": TokenType.RENAME,
 811        "REPLACE": TokenType.REPLACE,
 812        "RETURNING": TokenType.RETURNING,
 813        "REFERENCES": TokenType.REFERENCES,
 814        "RIGHT": TokenType.RIGHT,
 815        "RLIKE": TokenType.RLIKE,
 816        "ROLLBACK": TokenType.ROLLBACK,
 817        "ROLLUP": TokenType.ROLLUP,
 818        "ROW": TokenType.ROW,
 819        "ROWS": TokenType.ROWS,
 820        "SCHEMA": TokenType.SCHEMA,
 821        "SELECT": TokenType.SELECT,
 822        "SEMI": TokenType.SEMI,
 823        "SET": TokenType.SET,
 824        "SETTINGS": TokenType.SETTINGS,
 825        "SHOW": TokenType.SHOW,
 826        "SIMILAR TO": TokenType.SIMILAR_TO,
 827        "SOME": TokenType.SOME,
 828        "SORT BY": TokenType.SORT_BY,
 829        "START WITH": TokenType.START_WITH,
 830        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 831        "TABLE": TokenType.TABLE,
 832        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 833        "TEMP": TokenType.TEMPORARY,
 834        "TEMPORARY": TokenType.TEMPORARY,
 835        "THEN": TokenType.THEN,
 836        "TRUE": TokenType.TRUE,
 837        "TRUNCATE": TokenType.TRUNCATE,
 838        "UNION": TokenType.UNION,
 839        "UNKNOWN": TokenType.UNKNOWN,
 840        "UNNEST": TokenType.UNNEST,
 841        "UNPIVOT": TokenType.UNPIVOT,
 842        "UPDATE": TokenType.UPDATE,
 843        "USE": TokenType.USE,
 844        "USING": TokenType.USING,
 845        "UUID": TokenType.UUID,
 846        "VALUES": TokenType.VALUES,
 847        "VIEW": TokenType.VIEW,
 848        "VOLATILE": TokenType.VOLATILE,
 849        "WHEN": TokenType.WHEN,
 850        "WHERE": TokenType.WHERE,
 851        "WINDOW": TokenType.WINDOW,
 852        "WITH": TokenType.WITH,
 853        "APPLY": TokenType.APPLY,
 854        "ARRAY": TokenType.ARRAY,
 855        "BIT": TokenType.BIT,
 856        "BOOL": TokenType.BOOLEAN,
 857        "BOOLEAN": TokenType.BOOLEAN,
 858        "BYTE": TokenType.TINYINT,
 859        "MEDIUMINT": TokenType.MEDIUMINT,
 860        "INT1": TokenType.TINYINT,
 861        "TINYINT": TokenType.TINYINT,
 862        "INT16": TokenType.SMALLINT,
 863        "SHORT": TokenType.SMALLINT,
 864        "SMALLINT": TokenType.SMALLINT,
 865        "HUGEINT": TokenType.INT128,
 866        "UHUGEINT": TokenType.UINT128,
 867        "INT2": TokenType.SMALLINT,
 868        "INTEGER": TokenType.INT,
 869        "INT": TokenType.INT,
 870        "INT4": TokenType.INT,
 871        "INT32": TokenType.INT,
 872        "INT64": TokenType.BIGINT,
 873        "INT128": TokenType.INT128,
 874        "INT256": TokenType.INT256,
 875        "LONG": TokenType.BIGINT,
 876        "BIGINT": TokenType.BIGINT,
 877        "INT8": TokenType.TINYINT,
 878        "UINT": TokenType.UINT,
 879        "UINT128": TokenType.UINT128,
 880        "UINT256": TokenType.UINT256,
 881        "DEC": TokenType.DECIMAL,
 882        "DECIMAL": TokenType.DECIMAL,
 883        "DECIMAL32": TokenType.DECIMAL32,
 884        "DECIMAL64": TokenType.DECIMAL64,
 885        "DECIMAL128": TokenType.DECIMAL128,
 886        "DECIMAL256": TokenType.DECIMAL256,
 887        "BIGDECIMAL": TokenType.BIGDECIMAL,
 888        "BIGNUMERIC": TokenType.BIGDECIMAL,
 889        "LIST": TokenType.LIST,
 890        "MAP": TokenType.MAP,
 891        "NULLABLE": TokenType.NULLABLE,
 892        "NUMBER": TokenType.DECIMAL,
 893        "NUMERIC": TokenType.DECIMAL,
 894        "FIXED": TokenType.DECIMAL,
 895        "REAL": TokenType.FLOAT,
 896        "FLOAT": TokenType.FLOAT,
 897        "FLOAT4": TokenType.FLOAT,
 898        "FLOAT8": TokenType.DOUBLE,
 899        "DOUBLE": TokenType.DOUBLE,
 900        "DOUBLE PRECISION": TokenType.DOUBLE,
 901        "JSON": TokenType.JSON,
 902        "JSONB": TokenType.JSONB,
 903        "CHAR": TokenType.CHAR,
 904        "CHARACTER": TokenType.CHAR,
 905        "NCHAR": TokenType.NCHAR,
 906        "VARCHAR": TokenType.VARCHAR,
 907        "VARCHAR2": TokenType.VARCHAR,
 908        "NVARCHAR": TokenType.NVARCHAR,
 909        "NVARCHAR2": TokenType.NVARCHAR,
 910        "BPCHAR": TokenType.BPCHAR,
 911        "STR": TokenType.TEXT,
 912        "STRING": TokenType.TEXT,
 913        "TEXT": TokenType.TEXT,
 914        "LONGTEXT": TokenType.LONGTEXT,
 915        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 916        "TINYTEXT": TokenType.TINYTEXT,
 917        "CLOB": TokenType.TEXT,
 918        "LONGVARCHAR": TokenType.TEXT,
 919        "BINARY": TokenType.BINARY,
 920        "BLOB": TokenType.VARBINARY,
 921        "LONGBLOB": TokenType.LONGBLOB,
 922        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 923        "TINYBLOB": TokenType.TINYBLOB,
 924        "BYTEA": TokenType.VARBINARY,
 925        "VARBINARY": TokenType.VARBINARY,
 926        "TIME": TokenType.TIME,
 927        "TIMETZ": TokenType.TIMETZ,
 928        "TIMESTAMP": TokenType.TIMESTAMP,
 929        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 930        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 931        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 932        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 933        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 934        "DATE": TokenType.DATE,
 935        "DATETIME": TokenType.DATETIME,
 936        "INT4RANGE": TokenType.INT4RANGE,
 937        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 938        "INT8RANGE": TokenType.INT8RANGE,
 939        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 940        "NUMRANGE": TokenType.NUMRANGE,
 941        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 942        "TSRANGE": TokenType.TSRANGE,
 943        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 944        "TSTZRANGE": TokenType.TSTZRANGE,
 945        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 946        "DATERANGE": TokenType.DATERANGE,
 947        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 948        "UNIQUE": TokenType.UNIQUE,
 949        "VECTOR": TokenType.VECTOR,
 950        "STRUCT": TokenType.STRUCT,
 951        "SEQUENCE": TokenType.SEQUENCE,
 952        "VARIANT": TokenType.VARIANT,
 953        "ALTER": TokenType.ALTER,
 954        "ANALYZE": TokenType.ANALYZE,
 955        "CALL": TokenType.COMMAND,
 956        "COMMENT": TokenType.COMMENT,
 957        "EXPLAIN": TokenType.COMMAND,
 958        "GRANT": TokenType.GRANT,
 959        "OPTIMIZE": TokenType.COMMAND,
 960        "PREPARE": TokenType.COMMAND,
 961        "VACUUM": TokenType.COMMAND,
 962        "USER-DEFINED": TokenType.USERDEFINED,
 963        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 964        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 965    }
 966
 967    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 968        " ": TokenType.SPACE,
 969        "\t": TokenType.SPACE,
 970        "\n": TokenType.BREAK,
 971        "\r": TokenType.BREAK,
 972    }
 973
 974    COMMANDS = {
 975        TokenType.COMMAND,
 976        TokenType.EXECUTE,
 977        TokenType.FETCH,
 978        TokenType.SHOW,
 979        TokenType.RENAME,
 980    }
 981
 982    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 983
 984    # Handle numeric literals like in hive (3L = BIGINT)
 985    NUMERIC_LITERALS: t.Dict[str, str] = {}
 986
 987    COMMENTS = ["--", ("/*", "*/")]
 988
 989    __slots__ = (
 990        "sql",
 991        "size",
 992        "tokens",
 993        "dialect",
 994        "use_rs_tokenizer",
 995        "_start",
 996        "_current",
 997        "_line",
 998        "_col",
 999        "_comments",
1000        "_char",
1001        "_end",
1002        "_peek",
1003        "_prev_token_line",
1004        "_rs_dialect_settings",
1005    )
1006
1007    def __init__(
1008        self, dialect: DialectType = None, use_rs_tokenizer: t.Optional[bool] = None
1009    ) -> None:
1010        from sqlglot.dialects import Dialect
1011
1012        self.dialect = Dialect.get_or_raise(dialect)
1013
1014        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1015        self.use_rs_tokenizer = (
1016            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1017        )
1018
1019        if self.use_rs_tokenizer:
1020            self._rs_dialect_settings = RsTokenizerDialectSettings(
1021                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1022                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1023                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1024            )
1025
1026        self.reset()
1027
1028    def reset(self) -> None:
1029        self.sql = ""
1030        self.size = 0
1031        self.tokens: t.List[Token] = []
1032        self._start = 0
1033        self._current = 0
1034        self._line = 1
1035        self._col = 0
1036        self._comments: t.List[str] = []
1037
1038        self._char = ""
1039        self._end = False
1040        self._peek = ""
1041        self._prev_token_line = -1
1042
1043    def tokenize(self, sql: str) -> t.List[Token]:
1044        """Returns a list of tokens corresponding to the SQL string `sql`."""
1045        if self.use_rs_tokenizer:
1046            return self.tokenize_rs(sql)
1047
1048        self.reset()
1049        self.sql = sql
1050        self.size = len(sql)
1051
1052        try:
1053            self._scan()
1054        except Exception as e:
1055            start = max(self._current - 50, 0)
1056            end = min(self._current + 50, self.size - 1)
1057            context = self.sql[start:end]
1058            raise TokenError(f"Error tokenizing '{context}'") from e
1059
1060        return self.tokens
1061
1062    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1063        while self.size and not self._end:
1064            current = self._current
1065
1066            # Skip spaces here rather than iteratively calling advance() for performance reasons
1067            while current < self.size:
1068                char = self.sql[current]
1069
1070                if char.isspace() and (char == " " or char == "\t"):
1071                    current += 1
1072                else:
1073                    break
1074
1075            offset = current - self._current if current > self._current else 1
1076
1077            self._start = current
1078            self._advance(offset)
1079
1080            if not self._char.isspace():
1081                if self._char.isdigit():
1082                    self._scan_number()
1083                elif self._char in self._IDENTIFIERS:
1084                    self._scan_identifier(self._IDENTIFIERS[self._char])
1085                else:
1086                    self._scan_keywords()
1087
1088            if until and until():
1089                break
1090
1091        if self.tokens and self._comments:
1092            self.tokens[-1].comments.extend(self._comments)
1093
1094    def _chars(self, size: int) -> str:
1095        if size == 1:
1096            return self._char
1097
1098        start = self._current - 1
1099        end = start + size
1100
1101        return self.sql[start:end] if end <= self.size else ""
1102
1103    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1104        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1105            # Ensures we don't count an extra line if we get a \r\n line break sequence
1106            if not (self._char == "\r" and self._peek == "\n"):
1107                self._col = 1
1108                self._line += 1
1109        else:
1110            self._col += i
1111
1112        self._current += i
1113        self._end = self._current >= self.size
1114        self._char = self.sql[self._current - 1]
1115        self._peek = "" if self._end else self.sql[self._current]
1116
1117        if alnum and self._char.isalnum():
1118            # Here we use local variables instead of attributes for better performance
1119            _col = self._col
1120            _current = self._current
1121            _end = self._end
1122            _peek = self._peek
1123
1124            while _peek.isalnum():
1125                _col += 1
1126                _current += 1
1127                _end = _current >= self.size
1128                _peek = "" if _end else self.sql[_current]
1129
1130            self._col = _col
1131            self._current = _current
1132            self._end = _end
1133            self._peek = _peek
1134            self._char = self.sql[_current - 1]
1135
1136    @property
1137    def _text(self) -> str:
1138        return self.sql[self._start : self._current]
1139
1140    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1141        self._prev_token_line = self._line
1142
1143        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1144            self.tokens[-1].comments.extend(self._comments)
1145            self._comments = []
1146
1147        self.tokens.append(
1148            Token(
1149                token_type,
1150                text=self._text if text is None else text,
1151                line=self._line,
1152                col=self._col,
1153                start=self._start,
1154                end=self._current - 1,
1155                comments=self._comments,
1156            )
1157        )
1158        self._comments = []
1159
1160        # If we have either a semicolon or a begin token before the command's token, we'll parse
1161        # whatever follows the command's token as a string
1162        if (
1163            token_type in self.COMMANDS
1164            and self._peek != ";"
1165            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1166        ):
1167            start = self._current
1168            tokens = len(self.tokens)
1169            self._scan(lambda: self._peek == ";")
1170            self.tokens = self.tokens[:tokens]
1171            text = self.sql[start : self._current].strip()
1172            if text:
1173                self._add(TokenType.STRING, text)
1174
1175    def _scan_keywords(self) -> None:
1176        size = 0
1177        word = None
1178        chars = self._text
1179        char = chars
1180        prev_space = False
1181        skip = False
1182        trie = self._KEYWORD_TRIE
1183        single_token = char in self.SINGLE_TOKENS
1184
1185        while chars:
1186            if skip:
1187                result = TrieResult.PREFIX
1188            else:
1189                result, trie = in_trie(trie, char.upper())
1190
1191            if result == TrieResult.FAILED:
1192                break
1193            if result == TrieResult.EXISTS:
1194                word = chars
1195
1196            end = self._current + size
1197            size += 1
1198
1199            if end < self.size:
1200                char = self.sql[end]
1201                single_token = single_token or char in self.SINGLE_TOKENS
1202                is_space = char.isspace()
1203
1204                if not is_space or not prev_space:
1205                    if is_space:
1206                        char = " "
1207                    chars += char
1208                    prev_space = is_space
1209                    skip = False
1210                else:
1211                    skip = True
1212            else:
1213                char = ""
1214                break
1215
1216        if word:
1217            if self._scan_string(word):
1218                return
1219            if self._scan_comment(word):
1220                return
1221            if prev_space or single_token or not char:
1222                self._advance(size - 1)
1223                word = word.upper()
1224                self._add(self.KEYWORDS[word], text=word)
1225                return
1226
1227        if self._char in self.SINGLE_TOKENS:
1228            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1229            return
1230
1231        self._scan_var()
1232
1233    def _scan_comment(self, comment_start: str) -> bool:
1234        if comment_start not in self._COMMENTS:
1235            return False
1236
1237        comment_start_line = self._line
1238        comment_start_size = len(comment_start)
1239        comment_end = self._COMMENTS[comment_start]
1240
1241        if comment_end:
1242            # Skip the comment's start delimiter
1243            self._advance(comment_start_size)
1244
1245            comment_count = 1
1246            comment_end_size = len(comment_end)
1247
1248            while not self._end:
1249                if self._chars(comment_end_size) == comment_end:
1250                    comment_count -= 1
1251                    if not comment_count:
1252                        break
1253
1254                self._advance(alnum=True)
1255
1256                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1257                if (
1258                    self.NESTED_COMMENTS
1259                    and not self._end
1260                    and self._chars(comment_end_size) == comment_start
1261                ):
1262                    self._advance(comment_start_size)
1263                    comment_count += 1
1264
1265            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1266            self._advance(comment_end_size - 1)
1267        else:
1268            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1269                self._advance(alnum=True)
1270            self._comments.append(self._text[comment_start_size:])
1271
1272        if (
1273            comment_start == self.HINT_START
1274            and self.tokens
1275            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1276        ):
1277            self._add(TokenType.HINT)
1278
1279        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1280        # Multiple consecutive comments are preserved by appending them to the current comments list.
1281        if comment_start_line == self._prev_token_line:
1282            self.tokens[-1].comments.extend(self._comments)
1283            self._comments = []
1284            self._prev_token_line = self._line
1285
1286        return True
1287
1288    def _scan_number(self) -> None:
1289        if self._char == "0":
1290            peek = self._peek.upper()
1291            if peek == "B":
1292                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1293            elif peek == "X":
1294                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1295
1296        decimal = False
1297        scientific = 0
1298
1299        while True:
1300            if self._peek.isdigit():
1301                self._advance()
1302            elif self._peek == "." and not decimal:
1303                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1304                    return self._add(TokenType.NUMBER)
1305                decimal = True
1306                self._advance()
1307            elif self._peek in ("-", "+") and scientific == 1:
1308                scientific += 1
1309                self._advance()
1310            elif self._peek.upper() == "E" and not scientific:
1311                scientific += 1
1312                self._advance()
1313            elif self._peek.isidentifier():
1314                number_text = self._text
1315                literal = ""
1316
1317                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1318                    literal += self._peek
1319                    self._advance()
1320
1321                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1322
1323                if token_type:
1324                    self._add(TokenType.NUMBER, number_text)
1325                    self._add(TokenType.DCOLON, "::")
1326                    return self._add(token_type, literal)
1327                else:
1328                    replaced = literal.replace("_", "")
1329                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1330                        return self._add(TokenType.NUMBER, number_text + replaced)
1331                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1332                        return self._add(TokenType.VAR)
1333
1334                self._advance(-len(literal))
1335                return self._add(TokenType.NUMBER, number_text)
1336            else:
1337                return self._add(TokenType.NUMBER)
1338
1339    def _scan_bits(self) -> None:
1340        self._advance()
1341        value = self._extract_value()
1342        try:
1343            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1344            int(value, 2)
1345            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1346        except ValueError:
1347            self._add(TokenType.IDENTIFIER)
1348
1349    def _scan_hex(self) -> None:
1350        self._advance()
1351        value = self._extract_value()
1352        try:
1353            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1354            int(value, 16)
1355            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1356        except ValueError:
1357            self._add(TokenType.IDENTIFIER)
1358
1359    def _extract_value(self) -> str:
1360        while True:
1361            char = self._peek.strip()
1362            if char and char not in self.SINGLE_TOKENS:
1363                self._advance(alnum=True)
1364            else:
1365                break
1366
1367        return self._text
1368
1369    def _scan_string(self, start: str) -> bool:
1370        base = None
1371        token_type = TokenType.STRING
1372
1373        if start in self._QUOTES:
1374            end = self._QUOTES[start]
1375        elif start in self._FORMAT_STRINGS:
1376            end, token_type = self._FORMAT_STRINGS[start]
1377
1378            if token_type == TokenType.HEX_STRING:
1379                base = 16
1380            elif token_type == TokenType.BIT_STRING:
1381                base = 2
1382            elif token_type == TokenType.HEREDOC_STRING:
1383                self._advance()
1384
1385                if self._char == end:
1386                    tag = ""
1387                else:
1388                    tag = self._extract_string(
1389                        end,
1390                        raw_string=True,
1391                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1392                    )
1393
1394                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1395                    if not self._end:
1396                        self._advance(-1)
1397
1398                    self._advance(-len(tag))
1399                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1400                    return True
1401
1402                end = f"{start}{tag}{end}"
1403        else:
1404            return False
1405
1406        self._advance(len(start))
1407        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1408
1409        if base:
1410            try:
1411                int(text, base)
1412            except Exception:
1413                raise TokenError(
1414                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1415                )
1416
1417        self._add(token_type, text)
1418        return True
1419
1420    def _scan_identifier(self, identifier_end: str) -> None:
1421        self._advance()
1422        text = self._extract_string(
1423            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1424        )
1425        self._add(TokenType.IDENTIFIER, text)
1426
1427    def _scan_var(self) -> None:
1428        while True:
1429            char = self._peek.strip()
1430            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1431                self._advance(alnum=True)
1432            else:
1433                break
1434
1435        self._add(
1436            TokenType.VAR
1437            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1438            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1439        )
1440
1441    def _extract_string(
1442        self,
1443        delimiter: str,
1444        escapes: t.Optional[t.Set[str]] = None,
1445        raw_string: bool = False,
1446        raise_unmatched: bool = True,
1447    ) -> str:
1448        text = ""
1449        delim_size = len(delimiter)
1450        escapes = self._STRING_ESCAPES if escapes is None else escapes
1451
1452        while True:
1453            if (
1454                not raw_string
1455                and self.dialect.UNESCAPED_SEQUENCES
1456                and self._peek
1457                and self._char in self.STRING_ESCAPES
1458            ):
1459                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1460                if unescaped_sequence:
1461                    self._advance(2)
1462                    text += unescaped_sequence
1463                    continue
1464            if (
1465                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1466                and self._char in escapes
1467                and (self._peek == delimiter or self._peek in escapes)
1468                and (self._char not in self._QUOTES or self._char == self._peek)
1469            ):
1470                if self._peek == delimiter:
1471                    text += self._peek
1472                else:
1473                    text += self._char + self._peek
1474
1475                if self._current + 1 < self.size:
1476                    self._advance(2)
1477                else:
1478                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1479            else:
1480                if self._chars(delim_size) == delimiter:
1481                    if delim_size > 1:
1482                        self._advance(delim_size - 1)
1483                    break
1484
1485                if self._end:
1486                    if not raise_unmatched:
1487                        return text + self._char
1488
1489                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1490
1491                current = self._current - 1
1492                self._advance(alnum=True)
1493                text += self.sql[current : self._current - 1]
1494
1495        return text
1496
1497    def tokenize_rs(self, sql: str) -> t.List[Token]:
1498        if not self._RS_TOKENIZER:
1499            raise SqlglotError("Rust tokenizer is not available")
1500
1501        try:
1502            tokens = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1503            for token in tokens:
1504                token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1505            return tokens
1506        except Exception as e:
1507            raise TokenError(str(e))