sqlglot.tokens

   1from __future__ import annotations
   2
   3import os
   4import typing as t
   5from enum import auto
   6
   7from sqlglot.errors import SqlglotError, TokenError
   8from sqlglot.helper import AutoName
   9from sqlglot.trie import TrieResult, in_trie, new_trie
  10
  11if t.TYPE_CHECKING:
  12    from sqlglot.dialects.dialect import DialectType
  13
  14
  15try:
  16    from sqlglotrs import (  # type: ignore
  17        Tokenizer as RsTokenizer,
  18        TokenizerDialectSettings as RsTokenizerDialectSettings,
  19        TokenizerSettings as RsTokenizerSettings,
  20        TokenTypeSettings as RsTokenTypeSettings,
  21    )
  22
  23    USE_RS_TOKENIZER = os.environ.get("SQLGLOTRS_TOKENIZER", "1") == "1"
  24except ImportError:
  25    USE_RS_TOKENIZER = False
  26
  27
  28class TokenType(AutoName):
  29    L_PAREN = auto()
  30    R_PAREN = auto()
  31    L_BRACKET = auto()
  32    R_BRACKET = auto()
  33    L_BRACE = auto()
  34    R_BRACE = auto()
  35    COMMA = auto()
  36    DOT = auto()
  37    DASH = auto()
  38    PLUS = auto()
  39    COLON = auto()
  40    DOTCOLON = auto()
  41    DCOLON = auto()
  42    DCOLONDOLLAR = auto()
  43    DCOLONPERCENT = auto()
  44    DCOLONQMARK = auto()
  45    DQMARK = auto()
  46    SEMICOLON = auto()
  47    STAR = auto()
  48    BACKSLASH = auto()
  49    SLASH = auto()
  50    LT = auto()
  51    LTE = auto()
  52    GT = auto()
  53    GTE = auto()
  54    NOT = auto()
  55    EQ = auto()
  56    NEQ = auto()
  57    NULLSAFE_EQ = auto()
  58    COLON_EQ = auto()
  59    COLON_GT = auto()
  60    NCOLON_GT = auto()
  61    AND = auto()
  62    OR = auto()
  63    AMP = auto()
  64    DPIPE = auto()
  65    PIPE_GT = auto()
  66    PIPE = auto()
  67    PIPE_SLASH = auto()
  68    DPIPE_SLASH = auto()
  69    CARET = auto()
  70    CARET_AT = auto()
  71    TILDA = auto()
  72    ARROW = auto()
  73    DARROW = auto()
  74    FARROW = auto()
  75    HASH = auto()
  76    HASH_ARROW = auto()
  77    DHASH_ARROW = auto()
  78    LR_ARROW = auto()
  79    DAT = auto()
  80    LT_AT = auto()
  81    AT_GT = auto()
  82    DOLLAR = auto()
  83    PARAMETER = auto()
  84    SESSION = auto()
  85    SESSION_PARAMETER = auto()
  86    SESSION_USER = auto()
  87    DAMP = auto()
  88    AMP_LT = auto()
  89    AMP_GT = auto()
  90    ADJACENT = auto()
  91    XOR = auto()
  92    DSTAR = auto()
  93    QMARK_AMP = auto()
  94    QMARK_PIPE = auto()
  95    HASH_DASH = auto()
  96    EXCLAMATION = auto()
  97
  98    URI_START = auto()
  99
 100    BLOCK_START = auto()
 101    BLOCK_END = auto()
 102
 103    SPACE = auto()
 104    BREAK = auto()
 105
 106    STRING = auto()
 107    NUMBER = auto()
 108    IDENTIFIER = auto()
 109    DATABASE = auto()
 110    COLUMN = auto()
 111    COLUMN_DEF = auto()
 112    SCHEMA = auto()
 113    TABLE = auto()
 114    WAREHOUSE = auto()
 115    STAGE = auto()
 116    STREAMLIT = auto()
 117    VAR = auto()
 118    BIT_STRING = auto()
 119    HEX_STRING = auto()
 120    BYTE_STRING = auto()
 121    NATIONAL_STRING = auto()
 122    RAW_STRING = auto()
 123    HEREDOC_STRING = auto()
 124    UNICODE_STRING = auto()
 125
 126    # types
 127    BIT = auto()
 128    BOOLEAN = auto()
 129    TINYINT = auto()
 130    UTINYINT = auto()
 131    SMALLINT = auto()
 132    USMALLINT = auto()
 133    MEDIUMINT = auto()
 134    UMEDIUMINT = auto()
 135    INT = auto()
 136    UINT = auto()
 137    BIGINT = auto()
 138    UBIGINT = auto()
 139    BIGNUM = auto()  # unlimited precision int
 140    INT128 = auto()
 141    UINT128 = auto()
 142    INT256 = auto()
 143    UINT256 = auto()
 144    FLOAT = auto()
 145    DOUBLE = auto()
 146    UDOUBLE = auto()
 147    DECIMAL = auto()
 148    DECIMAL32 = auto()
 149    DECIMAL64 = auto()
 150    DECIMAL128 = auto()
 151    DECIMAL256 = auto()
 152    DECFLOAT = auto()
 153    UDECIMAL = auto()
 154    BIGDECIMAL = auto()
 155    CHAR = auto()
 156    NCHAR = auto()
 157    VARCHAR = auto()
 158    NVARCHAR = auto()
 159    BPCHAR = auto()
 160    TEXT = auto()
 161    MEDIUMTEXT = auto()
 162    LONGTEXT = auto()
 163    BLOB = auto()
 164    MEDIUMBLOB = auto()
 165    LONGBLOB = auto()
 166    TINYBLOB = auto()
 167    TINYTEXT = auto()
 168    NAME = auto()
 169    BINARY = auto()
 170    VARBINARY = auto()
 171    JSON = auto()
 172    JSONB = auto()
 173    TIME = auto()
 174    TIMETZ = auto()
 175    TIME_NS = auto()
 176    TIMESTAMP = auto()
 177    TIMESTAMPTZ = auto()
 178    TIMESTAMPLTZ = auto()
 179    TIMESTAMPNTZ = auto()
 180    TIMESTAMP_S = auto()
 181    TIMESTAMP_MS = auto()
 182    TIMESTAMP_NS = auto()
 183    DATETIME = auto()
 184    DATETIME2 = auto()
 185    DATETIME64 = auto()
 186    SMALLDATETIME = auto()
 187    DATE = auto()
 188    DATE32 = auto()
 189    INT4RANGE = auto()
 190    INT4MULTIRANGE = auto()
 191    INT8RANGE = auto()
 192    INT8MULTIRANGE = auto()
 193    NUMRANGE = auto()
 194    NUMMULTIRANGE = auto()
 195    TSRANGE = auto()
 196    TSMULTIRANGE = auto()
 197    TSTZRANGE = auto()
 198    TSTZMULTIRANGE = auto()
 199    DATERANGE = auto()
 200    DATEMULTIRANGE = auto()
 201    UUID = auto()
 202    GEOGRAPHY = auto()
 203    GEOGRAPHYPOINT = auto()
 204    NULLABLE = auto()
 205    GEOMETRY = auto()
 206    POINT = auto()
 207    RING = auto()
 208    LINESTRING = auto()
 209    LOCALTIME = auto()
 210    LOCALTIMESTAMP = auto()
 211    SYSTIMESTAMP = auto()
 212    MULTILINESTRING = auto()
 213    POLYGON = auto()
 214    MULTIPOLYGON = auto()
 215    HLLSKETCH = auto()
 216    HSTORE = auto()
 217    SUPER = auto()
 218    SERIAL = auto()
 219    SMALLSERIAL = auto()
 220    BIGSERIAL = auto()
 221    XML = auto()
 222    YEAR = auto()
 223    USERDEFINED = auto()
 224    MONEY = auto()
 225    SMALLMONEY = auto()
 226    ROWVERSION = auto()
 227    IMAGE = auto()
 228    VARIANT = auto()
 229    OBJECT = auto()
 230    INET = auto()
 231    IPADDRESS = auto()
 232    IPPREFIX = auto()
 233    IPV4 = auto()
 234    IPV6 = auto()
 235    ENUM = auto()
 236    ENUM8 = auto()
 237    ENUM16 = auto()
 238    FIXEDSTRING = auto()
 239    LOWCARDINALITY = auto()
 240    NESTED = auto()
 241    AGGREGATEFUNCTION = auto()
 242    SIMPLEAGGREGATEFUNCTION = auto()
 243    TDIGEST = auto()
 244    UNKNOWN = auto()
 245    VECTOR = auto()
 246    DYNAMIC = auto()
 247    VOID = auto()
 248
 249    # keywords
 250    ALIAS = auto()
 251    ALTER = auto()
 252    ALL = auto()
 253    ANTI = auto()
 254    ANY = auto()
 255    APPLY = auto()
 256    ARRAY = auto()
 257    ASC = auto()
 258    ASOF = auto()
 259    ATTACH = auto()
 260    AUTO_INCREMENT = auto()
 261    BEGIN = auto()
 262    BETWEEN = auto()
 263    BULK_COLLECT_INTO = auto()
 264    CACHE = auto()
 265    CASE = auto()
 266    CHARACTER_SET = auto()
 267    CLUSTER_BY = auto()
 268    COLLATE = auto()
 269    COMMAND = auto()
 270    COMMENT = auto()
 271    COMMIT = auto()
 272    CONNECT_BY = auto()
 273    CONSTRAINT = auto()
 274    COPY = auto()
 275    CREATE = auto()
 276    CROSS = auto()
 277    CUBE = auto()
 278    CURRENT_DATE = auto()
 279    CURRENT_DATETIME = auto()
 280    CURRENT_SCHEMA = auto()
 281    CURRENT_TIME = auto()
 282    CURRENT_TIMESTAMP = auto()
 283    CURRENT_USER = auto()
 284    CURRENT_ROLE = auto()
 285    CURRENT_CATALOG = auto()
 286    DECLARE = auto()
 287    DEFAULT = auto()
 288    DELETE = auto()
 289    DESC = auto()
 290    DESCRIBE = auto()
 291    DETACH = auto()
 292    DICTIONARY = auto()
 293    DISTINCT = auto()
 294    DISTRIBUTE_BY = auto()
 295    DIV = auto()
 296    DROP = auto()
 297    ELSE = auto()
 298    END = auto()
 299    ESCAPE = auto()
 300    EXCEPT = auto()
 301    EXECUTE = auto()
 302    EXISTS = auto()
 303    FALSE = auto()
 304    FETCH = auto()
 305    FILE = auto()
 306    FILE_FORMAT = auto()
 307    FILTER = auto()
 308    FINAL = auto()
 309    FIRST = auto()
 310    FOR = auto()
 311    FORCE = auto()
 312    FOREIGN_KEY = auto()
 313    FORMAT = auto()
 314    FROM = auto()
 315    FULL = auto()
 316    FUNCTION = auto()
 317    GET = auto()
 318    GLOB = auto()
 319    GLOBAL = auto()
 320    GRANT = auto()
 321    GROUP_BY = auto()
 322    GROUPING_SETS = auto()
 323    HAVING = auto()
 324    HINT = auto()
 325    IGNORE = auto()
 326    ILIKE = auto()
 327    IN = auto()
 328    INDEX = auto()
 329    INDEXED_BY = auto()
 330    INNER = auto()
 331    INSERT = auto()
 332    INSTALL = auto()
 333    INTERSECT = auto()
 334    INTERVAL = auto()
 335    INTO = auto()
 336    INTRODUCER = auto()
 337    IRLIKE = auto()
 338    IS = auto()
 339    ISNULL = auto()
 340    JOIN = auto()
 341    JOIN_MARKER = auto()
 342    KEEP = auto()
 343    KEY = auto()
 344    KILL = auto()
 345    LANGUAGE = auto()
 346    LATERAL = auto()
 347    LEFT = auto()
 348    LIKE = auto()
 349    LIMIT = auto()
 350    LIST = auto()
 351    LOAD = auto()
 352    LOCK = auto()
 353    MAP = auto()
 354    MATCH = auto()
 355    MATCH_CONDITION = auto()
 356    MATCH_RECOGNIZE = auto()
 357    MEMBER_OF = auto()
 358    MERGE = auto()
 359    MOD = auto()
 360    MODEL = auto()
 361    NATURAL = auto()
 362    NEXT = auto()
 363    NOTHING = auto()
 364    NOTNULL = auto()
 365    NULL = auto()
 366    OBJECT_IDENTIFIER = auto()
 367    OFFSET = auto()
 368    ON = auto()
 369    ONLY = auto()
 370    OPERATOR = auto()
 371    ORDER_BY = auto()
 372    ORDER_SIBLINGS_BY = auto()
 373    ORDERED = auto()
 374    ORDINALITY = auto()
 375    OUT = auto()
 376    OUTER = auto()
 377    OVER = auto()
 378    OVERLAPS = auto()
 379    OVERWRITE = auto()
 380    PARTITION = auto()
 381    PARTITION_BY = auto()
 382    PERCENT = auto()
 383    PIVOT = auto()
 384    PLACEHOLDER = auto()
 385    POSITIONAL = auto()
 386    PRAGMA = auto()
 387    PREWHERE = auto()
 388    PRIMARY_KEY = auto()
 389    PROCEDURE = auto()
 390    PROPERTIES = auto()
 391    PSEUDO_TYPE = auto()
 392    PUT = auto()
 393    QUALIFY = auto()
 394    QUOTE = auto()
 395    QDCOLON = auto()
 396    RANGE = auto()
 397    RECURSIVE = auto()
 398    REFRESH = auto()
 399    RENAME = auto()
 400    REPLACE = auto()
 401    RETURNING = auto()
 402    REVOKE = auto()
 403    REFERENCES = auto()
 404    RIGHT = auto()
 405    RLIKE = auto()
 406    ROLLBACK = auto()
 407    ROLLUP = auto()
 408    ROW = auto()
 409    ROWS = auto()
 410    SELECT = auto()
 411    SEMI = auto()
 412    SEPARATOR = auto()
 413    SEQUENCE = auto()
 414    SERDE_PROPERTIES = auto()
 415    SET = auto()
 416    SETTINGS = auto()
 417    SHOW = auto()
 418    SIMILAR_TO = auto()
 419    SOME = auto()
 420    SORT_BY = auto()
 421    SOUNDS_LIKE = auto()
 422    START_WITH = auto()
 423    STORAGE_INTEGRATION = auto()
 424    STRAIGHT_JOIN = auto()
 425    STRUCT = auto()
 426    SUMMARIZE = auto()
 427    TABLE_SAMPLE = auto()
 428    TAG = auto()
 429    TEMPORARY = auto()
 430    TOP = auto()
 431    THEN = auto()
 432    TRUE = auto()
 433    TRUNCATE = auto()
 434    UNCACHE = auto()
 435    UNION = auto()
 436    UNNEST = auto()
 437    UNPIVOT = auto()
 438    UPDATE = auto()
 439    USE = auto()
 440    USING = auto()
 441    VALUES = auto()
 442    VIEW = auto()
 443    SEMANTIC_VIEW = auto()
 444    VOLATILE = auto()
 445    WHEN = auto()
 446    WHERE = auto()
 447    WINDOW = auto()
 448    WITH = auto()
 449    UNIQUE = auto()
 450    UTC_DATE = auto()
 451    UTC_TIME = auto()
 452    UTC_TIMESTAMP = auto()
 453    VERSION_SNAPSHOT = auto()
 454    TIMESTAMP_SNAPSHOT = auto()
 455    OPTION = auto()
 456    SINK = auto()
 457    SOURCE = auto()
 458    ANALYZE = auto()
 459    NAMESPACE = auto()
 460    EXPORT = auto()
 461
 462    # sentinel
 463    HIVE_TOKEN_STREAM = auto()
 464
 465
 466_ALL_TOKEN_TYPES = list(TokenType)
 467_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 468
 469
 470class Token:
 471    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 472
 473    @classmethod
 474    def number(cls, number: int) -> Token:
 475        """Returns a NUMBER token with `number` as its text."""
 476        return cls(TokenType.NUMBER, str(number))
 477
 478    @classmethod
 479    def string(cls, string: str) -> Token:
 480        """Returns a STRING token with `string` as its text."""
 481        return cls(TokenType.STRING, string)
 482
 483    @classmethod
 484    def identifier(cls, identifier: str) -> Token:
 485        """Returns an IDENTIFIER token with `identifier` as its text."""
 486        return cls(TokenType.IDENTIFIER, identifier)
 487
 488    @classmethod
 489    def var(cls, var: str) -> Token:
 490        """Returns an VAR token with `var` as its text."""
 491        return cls(TokenType.VAR, var)
 492
 493    def __init__(
 494        self,
 495        token_type: TokenType,
 496        text: str,
 497        line: int = 1,
 498        col: int = 1,
 499        start: int = 0,
 500        end: int = 0,
 501        comments: t.Optional[t.List[str]] = None,
 502    ) -> None:
 503        """Token initializer.
 504
 505        Args:
 506            token_type: The TokenType Enum.
 507            text: The text of the token.
 508            line: The line that the token ends on.
 509            col: The column that the token ends on.
 510            start: The start index of the token.
 511            end: The ending index of the token.
 512            comments: The comments to attach to the token.
 513        """
 514        self.token_type = token_type
 515        self.text = text
 516        self.line = line
 517        self.col = col
 518        self.start = start
 519        self.end = end
 520        self.comments = [] if comments is None else comments
 521
 522    def __repr__(self) -> str:
 523        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 524        return f"<Token {attributes}>"
 525
 526
 527class _Tokenizer(type):
 528    def __new__(cls, clsname, bases, attrs):
 529        klass = super().__new__(cls, clsname, bases, attrs)
 530
 531        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 532            return dict(
 533                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 534            )
 535
 536        def _quotes_to_format(
 537            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 538        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 539            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 540
 541        klass._QUOTES = _convert_quotes(klass.QUOTES)
 542        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 543
 544        klass._FORMAT_STRINGS = {
 545            **{
 546                p + s: (e, TokenType.NATIONAL_STRING)
 547                for s, e in klass._QUOTES.items()
 548                for p in ("n", "N")
 549            },
 550            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 551            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 552            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 553            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 554            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 555            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 556        }
 557
 558        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 559        klass._ESCAPE_FOLLOW_CHARS = set(klass.ESCAPE_FOLLOW_CHARS)
 560        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 561        klass._COMMENTS = {
 562            **dict(
 563                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 564                for comment in klass.COMMENTS
 565            ),
 566            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 567        }
 568        if klass.HINT_START in klass.KEYWORDS:
 569            klass._COMMENTS[klass.HINT_START] = "*/"
 570
 571        klass._KEYWORD_TRIE = new_trie(
 572            key.upper()
 573            for key in (
 574                *klass.KEYWORDS,
 575                *klass._COMMENTS,
 576                *klass._QUOTES,
 577                *klass._FORMAT_STRINGS,
 578            )
 579            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 580        )
 581
 582        if USE_RS_TOKENIZER:
 583            settings = RsTokenizerSettings(
 584                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 585                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 586                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 587                numeric_literals=klass.NUMERIC_LITERALS,
 588                identifiers=klass._IDENTIFIERS,
 589                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 590                string_escapes=klass._STRING_ESCAPES,
 591                quotes=klass._QUOTES,
 592                format_strings={
 593                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 594                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 595                },
 596                has_bit_strings=bool(klass.BIT_STRINGS),
 597                has_hex_strings=bool(klass.HEX_STRINGS),
 598                comments=klass._COMMENTS,
 599                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 600                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 601                command_prefix_tokens={
 602                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 603                },
 604                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 605                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 606                nested_comments=klass.NESTED_COMMENTS,
 607                hint_start=klass.HINT_START,
 608                tokens_preceding_hint={
 609                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 610                },
 611                escape_follow_chars=klass._ESCAPE_FOLLOW_CHARS,
 612            )
 613            token_types = RsTokenTypeSettings(
 614                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 615                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 616                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 617                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 618                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 619                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 620                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 621                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 622                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 623                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 624                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 625                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 626                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 627                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 628            )
 629            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 630        else:
 631            klass._RS_TOKENIZER = None
 632
 633        return klass
 634
 635
 636class Tokenizer(metaclass=_Tokenizer):
 637    SINGLE_TOKENS = {
 638        "(": TokenType.L_PAREN,
 639        ")": TokenType.R_PAREN,
 640        "[": TokenType.L_BRACKET,
 641        "]": TokenType.R_BRACKET,
 642        "{": TokenType.L_BRACE,
 643        "}": TokenType.R_BRACE,
 644        "&": TokenType.AMP,
 645        "^": TokenType.CARET,
 646        ":": TokenType.COLON,
 647        ",": TokenType.COMMA,
 648        ".": TokenType.DOT,
 649        "-": TokenType.DASH,
 650        "=": TokenType.EQ,
 651        ">": TokenType.GT,
 652        "<": TokenType.LT,
 653        "%": TokenType.MOD,
 654        "!": TokenType.NOT,
 655        "|": TokenType.PIPE,
 656        "+": TokenType.PLUS,
 657        ";": TokenType.SEMICOLON,
 658        "/": TokenType.SLASH,
 659        "\\": TokenType.BACKSLASH,
 660        "*": TokenType.STAR,
 661        "~": TokenType.TILDA,
 662        "?": TokenType.PLACEHOLDER,
 663        "@": TokenType.PARAMETER,
 664        "#": TokenType.HASH,
 665        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 666        "'": TokenType.UNKNOWN,
 667        "`": TokenType.UNKNOWN,
 668        '"': TokenType.UNKNOWN,
 669    }
 670
 671    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 672    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 673    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 674    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 675    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 676    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 677    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 678    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 679    STRING_ESCAPES = ["'"]
 680    VAR_SINGLE_TOKENS: t.Set[str] = set()
 681    ESCAPE_FOLLOW_CHARS: t.List[str] = []
 682
 683    # The strings in this list can always be used as escapes, regardless of the surrounding
 684    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 685    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 686    IDENTIFIER_ESCAPES: t.List[str] = []
 687
 688    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 689    HEREDOC_TAG_IS_IDENTIFIER = False
 690
 691    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 692    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 693
 694    # Whether string escape characters function as such when placed within raw strings
 695    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 696
 697    NESTED_COMMENTS = True
 698
 699    HINT_START = "/*+"
 700
 701    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 702
 703    # Autofilled
 704    _COMMENTS: t.Dict[str, str] = {}
 705    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 706    _IDENTIFIERS: t.Dict[str, str] = {}
 707    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 708    _QUOTES: t.Dict[str, str] = {}
 709    _STRING_ESCAPES: t.Set[str] = set()
 710    _KEYWORD_TRIE: t.Dict = {}
 711    _RS_TOKENIZER: t.Optional[t.Any] = None
 712    _ESCAPE_FOLLOW_CHARS: t.Set[str] = set()
 713
 714    KEYWORDS: t.Dict[str, TokenType] = {
 715        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 716        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 717        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 718        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 719        HINT_START: TokenType.HINT,
 720        "&<": TokenType.AMP_LT,
 721        "&>": TokenType.AMP_GT,
 722        "==": TokenType.EQ,
 723        "::": TokenType.DCOLON,
 724        "?::": TokenType.QDCOLON,
 725        "||": TokenType.DPIPE,
 726        "|>": TokenType.PIPE_GT,
 727        ">=": TokenType.GTE,
 728        "<=": TokenType.LTE,
 729        "<>": TokenType.NEQ,
 730        "!=": TokenType.NEQ,
 731        ":=": TokenType.COLON_EQ,
 732        "<=>": TokenType.NULLSAFE_EQ,
 733        "->": TokenType.ARROW,
 734        "->>": TokenType.DARROW,
 735        "=>": TokenType.FARROW,
 736        "#>": TokenType.HASH_ARROW,
 737        "#>>": TokenType.DHASH_ARROW,
 738        "<->": TokenType.LR_ARROW,
 739        "&&": TokenType.DAMP,
 740        "??": TokenType.DQMARK,
 741        "~~~": TokenType.GLOB,
 742        "~~": TokenType.LIKE,
 743        "~~*": TokenType.ILIKE,
 744        "~*": TokenType.IRLIKE,
 745        "-|-": TokenType.ADJACENT,
 746        "ALL": TokenType.ALL,
 747        "AND": TokenType.AND,
 748        "ANTI": TokenType.ANTI,
 749        "ANY": TokenType.ANY,
 750        "ASC": TokenType.ASC,
 751        "AS": TokenType.ALIAS,
 752        "ASOF": TokenType.ASOF,
 753        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 754        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 755        "BEGIN": TokenType.BEGIN,
 756        "BETWEEN": TokenType.BETWEEN,
 757        "CACHE": TokenType.CACHE,
 758        "UNCACHE": TokenType.UNCACHE,
 759        "CASE": TokenType.CASE,
 760        "CHARACTER SET": TokenType.CHARACTER_SET,
 761        "CLUSTER BY": TokenType.CLUSTER_BY,
 762        "COLLATE": TokenType.COLLATE,
 763        "COLUMN": TokenType.COLUMN,
 764        "COMMIT": TokenType.COMMIT,
 765        "CONNECT BY": TokenType.CONNECT_BY,
 766        "CONSTRAINT": TokenType.CONSTRAINT,
 767        "COPY": TokenType.COPY,
 768        "CREATE": TokenType.CREATE,
 769        "CROSS": TokenType.CROSS,
 770        "CUBE": TokenType.CUBE,
 771        "CURRENT_DATE": TokenType.CURRENT_DATE,
 772        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 773        "CURRENT_TIME": TokenType.CURRENT_TIME,
 774        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 775        "CURRENT_USER": TokenType.CURRENT_USER,
 776        "CURRENT_CATALOG": TokenType.CURRENT_CATALOG,
 777        "DATABASE": TokenType.DATABASE,
 778        "DEFAULT": TokenType.DEFAULT,
 779        "DELETE": TokenType.DELETE,
 780        "DESC": TokenType.DESC,
 781        "DESCRIBE": TokenType.DESCRIBE,
 782        "DISTINCT": TokenType.DISTINCT,
 783        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 784        "DIV": TokenType.DIV,
 785        "DROP": TokenType.DROP,
 786        "ELSE": TokenType.ELSE,
 787        "END": TokenType.END,
 788        "ENUM": TokenType.ENUM,
 789        "ESCAPE": TokenType.ESCAPE,
 790        "EXCEPT": TokenType.EXCEPT,
 791        "EXECUTE": TokenType.EXECUTE,
 792        "EXISTS": TokenType.EXISTS,
 793        "FALSE": TokenType.FALSE,
 794        "FETCH": TokenType.FETCH,
 795        "FILTER": TokenType.FILTER,
 796        "FILE": TokenType.FILE,
 797        "FIRST": TokenType.FIRST,
 798        "FULL": TokenType.FULL,
 799        "FUNCTION": TokenType.FUNCTION,
 800        "FOR": TokenType.FOR,
 801        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 802        "FORMAT": TokenType.FORMAT,
 803        "FROM": TokenType.FROM,
 804        "GEOGRAPHY": TokenType.GEOGRAPHY,
 805        "GEOMETRY": TokenType.GEOMETRY,
 806        "GLOB": TokenType.GLOB,
 807        "GROUP BY": TokenType.GROUP_BY,
 808        "GROUPING SETS": TokenType.GROUPING_SETS,
 809        "HAVING": TokenType.HAVING,
 810        "ILIKE": TokenType.ILIKE,
 811        "IN": TokenType.IN,
 812        "INDEX": TokenType.INDEX,
 813        "INET": TokenType.INET,
 814        "INNER": TokenType.INNER,
 815        "INSERT": TokenType.INSERT,
 816        "INTERVAL": TokenType.INTERVAL,
 817        "INTERSECT": TokenType.INTERSECT,
 818        "INTO": TokenType.INTO,
 819        "IS": TokenType.IS,
 820        "ISNULL": TokenType.ISNULL,
 821        "JOIN": TokenType.JOIN,
 822        "KEEP": TokenType.KEEP,
 823        "KILL": TokenType.KILL,
 824        "LATERAL": TokenType.LATERAL,
 825        "LEFT": TokenType.LEFT,
 826        "LIKE": TokenType.LIKE,
 827        "LIMIT": TokenType.LIMIT,
 828        "LOAD": TokenType.LOAD,
 829        "LOCALTIME": TokenType.LOCALTIME,
 830        "LOCALTIMESTAMP": TokenType.LOCALTIMESTAMP,
 831        "LOCK": TokenType.LOCK,
 832        "MERGE": TokenType.MERGE,
 833        "NAMESPACE": TokenType.NAMESPACE,
 834        "NATURAL": TokenType.NATURAL,
 835        "NEXT": TokenType.NEXT,
 836        "NOT": TokenType.NOT,
 837        "NOTNULL": TokenType.NOTNULL,
 838        "NULL": TokenType.NULL,
 839        "OBJECT": TokenType.OBJECT,
 840        "OFFSET": TokenType.OFFSET,
 841        "ON": TokenType.ON,
 842        "OR": TokenType.OR,
 843        "XOR": TokenType.XOR,
 844        "ORDER BY": TokenType.ORDER_BY,
 845        "ORDINALITY": TokenType.ORDINALITY,
 846        "OUT": TokenType.OUT,
 847        "OUTER": TokenType.OUTER,
 848        "OVER": TokenType.OVER,
 849        "OVERLAPS": TokenType.OVERLAPS,
 850        "OVERWRITE": TokenType.OVERWRITE,
 851        "PARTITION": TokenType.PARTITION,
 852        "PARTITION BY": TokenType.PARTITION_BY,
 853        "PARTITIONED BY": TokenType.PARTITION_BY,
 854        "PARTITIONED_BY": TokenType.PARTITION_BY,
 855        "PERCENT": TokenType.PERCENT,
 856        "PIVOT": TokenType.PIVOT,
 857        "PRAGMA": TokenType.PRAGMA,
 858        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 859        "PROCEDURE": TokenType.PROCEDURE,
 860        "OPERATOR": TokenType.OPERATOR,
 861        "QUALIFY": TokenType.QUALIFY,
 862        "RANGE": TokenType.RANGE,
 863        "RECURSIVE": TokenType.RECURSIVE,
 864        "REGEXP": TokenType.RLIKE,
 865        "RENAME": TokenType.RENAME,
 866        "REPLACE": TokenType.REPLACE,
 867        "RETURNING": TokenType.RETURNING,
 868        "REFERENCES": TokenType.REFERENCES,
 869        "RIGHT": TokenType.RIGHT,
 870        "RLIKE": TokenType.RLIKE,
 871        "ROLLBACK": TokenType.ROLLBACK,
 872        "ROLLUP": TokenType.ROLLUP,
 873        "ROW": TokenType.ROW,
 874        "ROWS": TokenType.ROWS,
 875        "SCHEMA": TokenType.SCHEMA,
 876        "SELECT": TokenType.SELECT,
 877        "SEMI": TokenType.SEMI,
 878        "SESSION": TokenType.SESSION,
 879        "SESSION_USER": TokenType.SESSION_USER,
 880        "SET": TokenType.SET,
 881        "SETTINGS": TokenType.SETTINGS,
 882        "SHOW": TokenType.SHOW,
 883        "SIMILAR TO": TokenType.SIMILAR_TO,
 884        "SOME": TokenType.SOME,
 885        "SORT BY": TokenType.SORT_BY,
 886        "START WITH": TokenType.START_WITH,
 887        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 888        "TABLE": TokenType.TABLE,
 889        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 890        "TEMP": TokenType.TEMPORARY,
 891        "TEMPORARY": TokenType.TEMPORARY,
 892        "THEN": TokenType.THEN,
 893        "TRUE": TokenType.TRUE,
 894        "TRUNCATE": TokenType.TRUNCATE,
 895        "UNION": TokenType.UNION,
 896        "UNKNOWN": TokenType.UNKNOWN,
 897        "UNNEST": TokenType.UNNEST,
 898        "UNPIVOT": TokenType.UNPIVOT,
 899        "UPDATE": TokenType.UPDATE,
 900        "USE": TokenType.USE,
 901        "USING": TokenType.USING,
 902        "UUID": TokenType.UUID,
 903        "VALUES": TokenType.VALUES,
 904        "VIEW": TokenType.VIEW,
 905        "VOLATILE": TokenType.VOLATILE,
 906        "WHEN": TokenType.WHEN,
 907        "WHERE": TokenType.WHERE,
 908        "WINDOW": TokenType.WINDOW,
 909        "WITH": TokenType.WITH,
 910        "APPLY": TokenType.APPLY,
 911        "ARRAY": TokenType.ARRAY,
 912        "BIT": TokenType.BIT,
 913        "BOOL": TokenType.BOOLEAN,
 914        "BOOLEAN": TokenType.BOOLEAN,
 915        "BYTE": TokenType.TINYINT,
 916        "MEDIUMINT": TokenType.MEDIUMINT,
 917        "INT1": TokenType.TINYINT,
 918        "TINYINT": TokenType.TINYINT,
 919        "INT16": TokenType.SMALLINT,
 920        "SHORT": TokenType.SMALLINT,
 921        "SMALLINT": TokenType.SMALLINT,
 922        "HUGEINT": TokenType.INT128,
 923        "UHUGEINT": TokenType.UINT128,
 924        "INT2": TokenType.SMALLINT,
 925        "INTEGER": TokenType.INT,
 926        "INT": TokenType.INT,
 927        "INT4": TokenType.INT,
 928        "INT32": TokenType.INT,
 929        "INT64": TokenType.BIGINT,
 930        "INT128": TokenType.INT128,
 931        "INT256": TokenType.INT256,
 932        "LONG": TokenType.BIGINT,
 933        "BIGINT": TokenType.BIGINT,
 934        "INT8": TokenType.TINYINT,
 935        "UINT": TokenType.UINT,
 936        "UINT128": TokenType.UINT128,
 937        "UINT256": TokenType.UINT256,
 938        "DEC": TokenType.DECIMAL,
 939        "DECIMAL": TokenType.DECIMAL,
 940        "DECIMAL32": TokenType.DECIMAL32,
 941        "DECIMAL64": TokenType.DECIMAL64,
 942        "DECIMAL128": TokenType.DECIMAL128,
 943        "DECIMAL256": TokenType.DECIMAL256,
 944        "DECFLOAT": TokenType.DECFLOAT,
 945        "BIGDECIMAL": TokenType.BIGDECIMAL,
 946        "BIGNUMERIC": TokenType.BIGDECIMAL,
 947        "BIGNUM": TokenType.BIGNUM,
 948        "LIST": TokenType.LIST,
 949        "MAP": TokenType.MAP,
 950        "NULLABLE": TokenType.NULLABLE,
 951        "NUMBER": TokenType.DECIMAL,
 952        "NUMERIC": TokenType.DECIMAL,
 953        "FIXED": TokenType.DECIMAL,
 954        "REAL": TokenType.FLOAT,
 955        "FLOAT": TokenType.FLOAT,
 956        "FLOAT4": TokenType.FLOAT,
 957        "FLOAT8": TokenType.DOUBLE,
 958        "DOUBLE": TokenType.DOUBLE,
 959        "DOUBLE PRECISION": TokenType.DOUBLE,
 960        "JSON": TokenType.JSON,
 961        "JSONB": TokenType.JSONB,
 962        "CHAR": TokenType.CHAR,
 963        "CHARACTER": TokenType.CHAR,
 964        "CHAR VARYING": TokenType.VARCHAR,
 965        "CHARACTER VARYING": TokenType.VARCHAR,
 966        "NCHAR": TokenType.NCHAR,
 967        "VARCHAR": TokenType.VARCHAR,
 968        "VARCHAR2": TokenType.VARCHAR,
 969        "NVARCHAR": TokenType.NVARCHAR,
 970        "NVARCHAR2": TokenType.NVARCHAR,
 971        "BPCHAR": TokenType.BPCHAR,
 972        "STR": TokenType.TEXT,
 973        "STRING": TokenType.TEXT,
 974        "TEXT": TokenType.TEXT,
 975        "LONGTEXT": TokenType.LONGTEXT,
 976        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 977        "TINYTEXT": TokenType.TINYTEXT,
 978        "CLOB": TokenType.TEXT,
 979        "LONGVARCHAR": TokenType.TEXT,
 980        "BINARY": TokenType.BINARY,
 981        "BLOB": TokenType.VARBINARY,
 982        "LONGBLOB": TokenType.LONGBLOB,
 983        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 984        "TINYBLOB": TokenType.TINYBLOB,
 985        "BYTEA": TokenType.VARBINARY,
 986        "VARBINARY": TokenType.VARBINARY,
 987        "TIME": TokenType.TIME,
 988        "TIMETZ": TokenType.TIMETZ,
 989        "TIME_NS": TokenType.TIME_NS,
 990        "TIMESTAMP": TokenType.TIMESTAMP,
 991        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 992        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 993        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 994        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 995        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 996        "DATE": TokenType.DATE,
 997        "DATETIME": TokenType.DATETIME,
 998        "INT4RANGE": TokenType.INT4RANGE,
 999        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
1000        "INT8RANGE": TokenType.INT8RANGE,
1001        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
1002        "NUMRANGE": TokenType.NUMRANGE,
1003        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
1004        "TSRANGE": TokenType.TSRANGE,
1005        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
1006        "TSTZRANGE": TokenType.TSTZRANGE,
1007        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
1008        "DATERANGE": TokenType.DATERANGE,
1009        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
1010        "UNIQUE": TokenType.UNIQUE,
1011        "VECTOR": TokenType.VECTOR,
1012        "STRUCT": TokenType.STRUCT,
1013        "SEQUENCE": TokenType.SEQUENCE,
1014        "VARIANT": TokenType.VARIANT,
1015        "ALTER": TokenType.ALTER,
1016        "ANALYZE": TokenType.ANALYZE,
1017        "CALL": TokenType.COMMAND,
1018        "COMMENT": TokenType.COMMENT,
1019        "EXPLAIN": TokenType.COMMAND,
1020        "GRANT": TokenType.GRANT,
1021        "REVOKE": TokenType.REVOKE,
1022        "OPTIMIZE": TokenType.COMMAND,
1023        "PREPARE": TokenType.COMMAND,
1024        "VACUUM": TokenType.COMMAND,
1025        "USER-DEFINED": TokenType.USERDEFINED,
1026        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
1027        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
1028    }
1029
1030    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
1031        " ": TokenType.SPACE,
1032        "\t": TokenType.SPACE,
1033        "\n": TokenType.BREAK,
1034        "\r": TokenType.BREAK,
1035    }
1036
1037    COMMANDS = {
1038        TokenType.COMMAND,
1039        TokenType.EXECUTE,
1040        TokenType.FETCH,
1041        TokenType.SHOW,
1042        TokenType.RENAME,
1043    }
1044
1045    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
1046
1047    # Handle numeric literals like in hive (3L = BIGINT)
1048    NUMERIC_LITERALS: t.Dict[str, str] = {}
1049
1050    COMMENTS = ["--", ("/*", "*/")]
1051
1052    __slots__ = (
1053        "sql",
1054        "size",
1055        "tokens",
1056        "dialect",
1057        "use_rs_tokenizer",
1058        "_start",
1059        "_current",
1060        "_line",
1061        "_col",
1062        "_comments",
1063        "_char",
1064        "_end",
1065        "_peek",
1066        "_prev_token_line",
1067        "_rs_dialect_settings",
1068    )
1069
1070    def __init__(
1071        self,
1072        dialect: DialectType = None,
1073        use_rs_tokenizer: t.Optional[bool] = None,
1074        **opts: t.Any,
1075    ) -> None:
1076        from sqlglot.dialects import Dialect
1077
1078        self.dialect = Dialect.get_or_raise(dialect)
1079
1080        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1081        self.use_rs_tokenizer = (
1082            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1083        )
1084
1085        if self.use_rs_tokenizer:
1086            self._rs_dialect_settings = RsTokenizerDialectSettings(
1087                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1088                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1089                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1090            )
1091
1092        self.reset()
1093
1094    def reset(self) -> None:
1095        self.sql = ""
1096        self.size = 0
1097        self.tokens: t.List[Token] = []
1098        self._start = 0
1099        self._current = 0
1100        self._line = 1
1101        self._col = 0
1102        self._comments: t.List[str] = []
1103
1104        self._char = ""
1105        self._end = False
1106        self._peek = ""
1107        self._prev_token_line = -1
1108
1109    def tokenize(self, sql: str) -> t.List[Token]:
1110        """Returns a list of tokens corresponding to the SQL string `sql`."""
1111        if self.use_rs_tokenizer:
1112            return self.tokenize_rs(sql)
1113
1114        self.reset()
1115        self.sql = sql
1116        self.size = len(sql)
1117
1118        try:
1119            self._scan()
1120        except Exception as e:
1121            start = max(self._current - 50, 0)
1122            end = min(self._current + 50, self.size - 1)
1123            context = self.sql[start:end]
1124            raise TokenError(f"Error tokenizing '{context}'") from e
1125
1126        return self.tokens
1127
1128    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1129        while self.size and not self._end:
1130            current = self._current
1131
1132            # Skip spaces here rather than iteratively calling advance() for performance reasons
1133            while current < self.size:
1134                char = self.sql[current]
1135
1136                if char.isspace() and (char == " " or char == "\t"):
1137                    current += 1
1138                else:
1139                    break
1140
1141            offset = current - self._current if current > self._current else 1
1142
1143            self._start = current
1144            self._advance(offset)
1145
1146            if not self._char.isspace():
1147                if self._char.isdigit():
1148                    self._scan_number()
1149                elif self._char in self._IDENTIFIERS:
1150                    self._scan_identifier(self._IDENTIFIERS[self._char])
1151                else:
1152                    self._scan_keywords()
1153
1154            if until and until():
1155                break
1156
1157        if self.tokens and self._comments:
1158            self.tokens[-1].comments.extend(self._comments)
1159
1160    def _chars(self, size: int) -> str:
1161        if size == 1:
1162            return self._char
1163
1164        start = self._current - 1
1165        end = start + size
1166
1167        return self.sql[start:end] if end <= self.size else ""
1168
1169    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1170        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1171            # Ensures we don't count an extra line if we get a \r\n line break sequence
1172            if not (self._char == "\r" and self._peek == "\n"):
1173                self._col = i
1174                self._line += 1
1175        else:
1176            self._col += i
1177
1178        self._current += i
1179        self._end = self._current >= self.size
1180        self._char = self.sql[self._current - 1]
1181        self._peek = "" if self._end else self.sql[self._current]
1182
1183        if alnum and self._char.isalnum():
1184            # Here we use local variables instead of attributes for better performance
1185            _col = self._col
1186            _current = self._current
1187            _end = self._end
1188            _peek = self._peek
1189
1190            while _peek.isalnum():
1191                _col += 1
1192                _current += 1
1193                _end = _current >= self.size
1194                _peek = "" if _end else self.sql[_current]
1195
1196            self._col = _col
1197            self._current = _current
1198            self._end = _end
1199            self._peek = _peek
1200            self._char = self.sql[_current - 1]
1201
1202    @property
1203    def _text(self) -> str:
1204        return self.sql[self._start : self._current]
1205
1206    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1207        self._prev_token_line = self._line
1208
1209        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1210            self.tokens[-1].comments.extend(self._comments)
1211            self._comments = []
1212
1213        self.tokens.append(
1214            Token(
1215                token_type,
1216                text=self._text if text is None else text,
1217                line=self._line,
1218                col=self._col,
1219                start=self._start,
1220                end=self._current - 1,
1221                comments=self._comments,
1222            )
1223        )
1224        self._comments = []
1225
1226        # If we have either a semicolon or a begin token before the command's token, we'll parse
1227        # whatever follows the command's token as a string
1228        if (
1229            token_type in self.COMMANDS
1230            and self._peek != ";"
1231            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1232        ):
1233            start = self._current
1234            tokens = len(self.tokens)
1235            self._scan(lambda: self._peek == ";")
1236            self.tokens = self.tokens[:tokens]
1237            text = self.sql[start : self._current].strip()
1238            if text:
1239                self._add(TokenType.STRING, text)
1240
1241    def _scan_keywords(self) -> None:
1242        size = 0
1243        word = None
1244        chars = self._text
1245        char = chars
1246        prev_space = False
1247        skip = False
1248        trie = self._KEYWORD_TRIE
1249        single_token = char in self.SINGLE_TOKENS
1250
1251        while chars:
1252            if skip:
1253                result = TrieResult.PREFIX
1254            else:
1255                result, trie = in_trie(trie, char.upper())
1256
1257            if result == TrieResult.FAILED:
1258                break
1259            if result == TrieResult.EXISTS:
1260                word = chars
1261
1262            end = self._current + size
1263            size += 1
1264
1265            if end < self.size:
1266                char = self.sql[end]
1267                single_token = single_token or char in self.SINGLE_TOKENS
1268                is_space = char.isspace()
1269
1270                if not is_space or not prev_space:
1271                    if is_space:
1272                        char = " "
1273                    chars += char
1274                    prev_space = is_space
1275                    skip = False
1276                else:
1277                    skip = True
1278            else:
1279                char = ""
1280                break
1281
1282        if word:
1283            if self._scan_string(word):
1284                return
1285            if self._scan_comment(word):
1286                return
1287            if prev_space or single_token or not char:
1288                self._advance(size - 1)
1289                word = word.upper()
1290                self._add(self.KEYWORDS[word], text=word)
1291                return
1292
1293        if self._char in self.SINGLE_TOKENS:
1294            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1295            return
1296
1297        self._scan_var()
1298
1299    def _scan_comment(self, comment_start: str) -> bool:
1300        if comment_start not in self._COMMENTS:
1301            return False
1302
1303        comment_start_line = self._line
1304        comment_start_size = len(comment_start)
1305        comment_end = self._COMMENTS[comment_start]
1306
1307        if comment_end:
1308            # Skip the comment's start delimiter
1309            self._advance(comment_start_size)
1310
1311            comment_count = 1
1312            comment_end_size = len(comment_end)
1313
1314            while not self._end:
1315                if self._chars(comment_end_size) == comment_end:
1316                    comment_count -= 1
1317                    if not comment_count:
1318                        break
1319
1320                self._advance(alnum=True)
1321
1322                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1323                if (
1324                    self.NESTED_COMMENTS
1325                    and not self._end
1326                    and self._chars(comment_end_size) == comment_start
1327                ):
1328                    self._advance(comment_start_size)
1329                    comment_count += 1
1330
1331            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1332            self._advance(comment_end_size - 1)
1333        else:
1334            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1335                self._advance(alnum=True)
1336            self._comments.append(self._text[comment_start_size:])
1337
1338        if (
1339            comment_start == self.HINT_START
1340            and self.tokens
1341            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1342        ):
1343            self._add(TokenType.HINT)
1344
1345        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1346        # Multiple consecutive comments are preserved by appending them to the current comments list.
1347        if comment_start_line == self._prev_token_line:
1348            self.tokens[-1].comments.extend(self._comments)
1349            self._comments = []
1350            self._prev_token_line = self._line
1351
1352        return True
1353
1354    def _scan_number(self) -> None:
1355        if self._char == "0":
1356            peek = self._peek.upper()
1357            if peek == "B":
1358                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1359            elif peek == "X":
1360                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1361
1362        decimal = False
1363        scientific = 0
1364
1365        while True:
1366            if self._peek.isdigit():
1367                self._advance()
1368            elif self._peek == "." and not decimal:
1369                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1370                    return self._add(TokenType.NUMBER)
1371                decimal = True
1372                self._advance()
1373            elif self._peek in ("-", "+") and scientific == 1:
1374                # Only consume +/- if followed by a digit
1375                if self._current + 1 < self.size and self.sql[self._current + 1].isdigit():
1376                    scientific += 1
1377                    self._advance()
1378                else:
1379                    return self._add(TokenType.NUMBER)
1380            elif self._peek.upper() == "E" and not scientific:
1381                scientific += 1
1382                self._advance()
1383            elif self._peek == "_" and self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED:
1384                self._advance()
1385            elif self._peek.isidentifier():
1386                number_text = self._text
1387                literal = ""
1388
1389                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1390                    literal += self._peek
1391                    self._advance()
1392
1393                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1394
1395                if token_type:
1396                    self._add(TokenType.NUMBER, number_text)
1397                    self._add(TokenType.DCOLON, "::")
1398                    return self._add(token_type, literal)
1399                elif self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1400                    return self._add(TokenType.VAR)
1401
1402                self._advance(-len(literal))
1403                return self._add(TokenType.NUMBER, number_text)
1404            else:
1405                return self._add(TokenType.NUMBER)
1406
1407    def _scan_bits(self) -> None:
1408        self._advance()
1409        value = self._extract_value()
1410        try:
1411            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1412            int(value, 2)
1413            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1414        except ValueError:
1415            self._add(TokenType.IDENTIFIER)
1416
1417    def _scan_hex(self) -> None:
1418        self._advance()
1419        value = self._extract_value()
1420        try:
1421            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1422            int(value, 16)
1423            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1424        except ValueError:
1425            self._add(TokenType.IDENTIFIER)
1426
1427    def _extract_value(self) -> str:
1428        while True:
1429            char = self._peek.strip()
1430            if char and char not in self.SINGLE_TOKENS:
1431                self._advance(alnum=True)
1432            else:
1433                break
1434
1435        return self._text
1436
1437    def _scan_string(self, start: str) -> bool:
1438        base = None
1439        token_type = TokenType.STRING
1440
1441        if start in self._QUOTES:
1442            end = self._QUOTES[start]
1443        elif start in self._FORMAT_STRINGS:
1444            end, token_type = self._FORMAT_STRINGS[start]
1445
1446            if token_type == TokenType.HEX_STRING:
1447                base = 16
1448            elif token_type == TokenType.BIT_STRING:
1449                base = 2
1450            elif token_type == TokenType.HEREDOC_STRING:
1451                self._advance()
1452
1453                if self._char == end:
1454                    tag = ""
1455                else:
1456                    tag = self._extract_string(
1457                        end,
1458                        raw_string=True,
1459                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1460                    )
1461
1462                if (
1463                    tag
1464                    and self.HEREDOC_TAG_IS_IDENTIFIER
1465                    and (self._end or tag.isdigit() or any(c.isspace() for c in tag))
1466                ):
1467                    if not self._end:
1468                        self._advance(-1)
1469
1470                    self._advance(-len(tag))
1471                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1472                    return True
1473
1474                end = f"{start}{tag}{end}"
1475        else:
1476            return False
1477
1478        self._advance(len(start))
1479        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1480
1481        if base and text:
1482            try:
1483                int(text, base)
1484            except Exception:
1485                raise TokenError(
1486                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1487                )
1488
1489        self._add(token_type, text)
1490        return True
1491
1492    def _scan_identifier(self, identifier_end: str) -> None:
1493        self._advance()
1494        text = self._extract_string(
1495            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1496        )
1497        self._add(TokenType.IDENTIFIER, text)
1498
1499    def _scan_var(self) -> None:
1500        while True:
1501            char = self._peek.strip()
1502            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1503                self._advance(alnum=True)
1504            else:
1505                break
1506
1507        self._add(
1508            TokenType.VAR
1509            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1510            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1511        )
1512
1513    def _extract_string(
1514        self,
1515        delimiter: str,
1516        escapes: t.Optional[t.Set[str]] = None,
1517        raw_string: bool = False,
1518        raise_unmatched: bool = True,
1519    ) -> str:
1520        text = ""
1521        delim_size = len(delimiter)
1522        escapes = self._STRING_ESCAPES if escapes is None else escapes
1523
1524        while True:
1525            if (
1526                not raw_string
1527                and self.dialect.UNESCAPED_SEQUENCES
1528                and self._peek
1529                and self._char in self.STRING_ESCAPES
1530            ):
1531                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1532                if unescaped_sequence:
1533                    self._advance(2)
1534                    text += unescaped_sequence
1535                    continue
1536
1537            is_valid_custom_escape = (
1538                self.ESCAPE_FOLLOW_CHARS
1539                and self._char == "\\"
1540                and self._peek not in self.ESCAPE_FOLLOW_CHARS
1541            )
1542
1543            if (
1544                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1545                and self._char in escapes
1546                and (self._peek == delimiter or self._peek in escapes or is_valid_custom_escape)
1547                and (self._char not in self._QUOTES or self._char == self._peek)
1548            ):
1549                if self._peek == delimiter:
1550                    text += self._peek
1551                elif is_valid_custom_escape and self._char != self._peek:
1552                    text += self._peek
1553                else:
1554                    text += self._char + self._peek
1555
1556                if self._current + 1 < self.size:
1557                    self._advance(2)
1558                else:
1559                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1560            else:
1561                if self._chars(delim_size) == delimiter:
1562                    if delim_size > 1:
1563                        self._advance(delim_size - 1)
1564                    break
1565
1566                if self._end:
1567                    if not raise_unmatched:
1568                        return text + self._char
1569
1570                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1571
1572                current = self._current - 1
1573                self._advance(alnum=True)
1574                text += self.sql[current : self._current - 1]
1575
1576        return text
1577
1578    def tokenize_rs(self, sql: str) -> t.List[Token]:
1579        if not self._RS_TOKENIZER:
1580            raise SqlglotError("Rust tokenizer is not available")
1581
1582        tokens, error_msg = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1583        for token in tokens:
1584            token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1585
1586        # Setting this here so partial token lists can be inspected even if there is a failure
1587        self.tokens = tokens
1588
1589        if error_msg is not None:
1590            raise TokenError(error_msg)
1591
1592        return tokens