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    TILDE = 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    INOUT = auto()
 377    OUTER = auto()
 378    OVER = auto()
 379    OVERLAPS = auto()
 380    OVERWRITE = auto()
 381    PARTITION = auto()
 382    PARTITION_BY = auto()
 383    PERCENT = auto()
 384    PIVOT = auto()
 385    PLACEHOLDER = auto()
 386    POSITIONAL = auto()
 387    PRAGMA = auto()
 388    PREWHERE = auto()
 389    PRIMARY_KEY = auto()
 390    PROCEDURE = auto()
 391    PROPERTIES = auto()
 392    PSEUDO_TYPE = auto()
 393    PUT = auto()
 394    QUALIFY = auto()
 395    QUOTE = auto()
 396    QDCOLON = auto()
 397    RANGE = auto()
 398    RECURSIVE = auto()
 399    REFRESH = auto()
 400    RENAME = auto()
 401    REPLACE = auto()
 402    RETURNING = auto()
 403    REVOKE = auto()
 404    REFERENCES = auto()
 405    RIGHT = auto()
 406    RLIKE = auto()
 407    ROLLBACK = auto()
 408    ROLLUP = auto()
 409    ROW = auto()
 410    ROWS = auto()
 411    SELECT = auto()
 412    SEMI = auto()
 413    SEPARATOR = auto()
 414    SEQUENCE = auto()
 415    SERDE_PROPERTIES = auto()
 416    SET = auto()
 417    SETTINGS = auto()
 418    SHOW = auto()
 419    SIMILAR_TO = auto()
 420    SOME = auto()
 421    SORT_BY = auto()
 422    SOUNDS_LIKE = auto()
 423    START_WITH = auto()
 424    STORAGE_INTEGRATION = auto()
 425    STRAIGHT_JOIN = auto()
 426    STRUCT = auto()
 427    SUMMARIZE = auto()
 428    TABLE_SAMPLE = auto()
 429    TAG = auto()
 430    TEMPORARY = auto()
 431    TOP = auto()
 432    THEN = auto()
 433    TRUE = auto()
 434    TRUNCATE = auto()
 435    UNCACHE = auto()
 436    UNION = auto()
 437    UNNEST = auto()
 438    UNPIVOT = auto()
 439    UPDATE = auto()
 440    USE = auto()
 441    USING = auto()
 442    VALUES = auto()
 443    VARIADIC = auto()
 444    VIEW = auto()
 445    SEMANTIC_VIEW = auto()
 446    VOLATILE = auto()
 447    WHEN = auto()
 448    WHERE = auto()
 449    WINDOW = auto()
 450    WITH = auto()
 451    UNIQUE = auto()
 452    UTC_DATE = auto()
 453    UTC_TIME = auto()
 454    UTC_TIMESTAMP = auto()
 455    VERSION_SNAPSHOT = auto()
 456    TIMESTAMP_SNAPSHOT = auto()
 457    OPTION = auto()
 458    SINK = auto()
 459    SOURCE = auto()
 460    ANALYZE = auto()
 461    NAMESPACE = auto()
 462    EXPORT = auto()
 463
 464    # sentinel
 465    HIVE_TOKEN_STREAM = auto()
 466
 467
 468_ALL_TOKEN_TYPES = list(TokenType)
 469_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 470
 471
 472class Token:
 473    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 474
 475    @classmethod
 476    def number(cls, number: int) -> Token:
 477        """Returns a NUMBER token with `number` as its text."""
 478        return cls(TokenType.NUMBER, str(number))
 479
 480    @classmethod
 481    def string(cls, string: str) -> Token:
 482        """Returns a STRING token with `string` as its text."""
 483        return cls(TokenType.STRING, string)
 484
 485    @classmethod
 486    def identifier(cls, identifier: str) -> Token:
 487        """Returns an IDENTIFIER token with `identifier` as its text."""
 488        return cls(TokenType.IDENTIFIER, identifier)
 489
 490    @classmethod
 491    def var(cls, var: str) -> Token:
 492        """Returns an VAR token with `var` as its text."""
 493        return cls(TokenType.VAR, var)
 494
 495    def __init__(
 496        self,
 497        token_type: TokenType,
 498        text: str,
 499        line: int = 1,
 500        col: int = 1,
 501        start: int = 0,
 502        end: int = 0,
 503        comments: t.Optional[t.List[str]] = None,
 504    ) -> None:
 505        """Token initializer.
 506
 507        Args:
 508            token_type: The TokenType Enum.
 509            text: The text of the token.
 510            line: The line that the token ends on.
 511            col: The column that the token ends on.
 512            start: The start index of the token.
 513            end: The ending index of the token.
 514            comments: The comments to attach to the token.
 515        """
 516        self.token_type = token_type
 517        self.text = text
 518        self.line = line
 519        self.col = col
 520        self.start = start
 521        self.end = end
 522        self.comments = [] if comments is None else comments
 523
 524    def __repr__(self) -> str:
 525        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 526        return f"<Token {attributes}>"
 527
 528
 529class _Tokenizer(type):
 530    def __new__(cls, clsname, bases, attrs):
 531        klass = super().__new__(cls, clsname, bases, attrs)
 532
 533        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 534            return dict(
 535                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 536            )
 537
 538        def _quotes_to_format(
 539            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 540        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 541            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 542
 543        klass._QUOTES = _convert_quotes(klass.QUOTES)
 544        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 545
 546        klass._FORMAT_STRINGS = {
 547            **{
 548                p + s: (e, TokenType.NATIONAL_STRING)
 549                for s, e in klass._QUOTES.items()
 550                for p in ("n", "N")
 551            },
 552            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 553            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 554            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 555            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 556            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 557            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 558        }
 559
 560        if "BYTE_STRING_ESCAPES" not in klass.__dict__:
 561            klass.BYTE_STRING_ESCAPES = klass.STRING_ESCAPES.copy()
 562
 563        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 564        klass._BYTE_STRING_ESCAPES = set(klass.BYTE_STRING_ESCAPES)
 565        klass._ESCAPE_FOLLOW_CHARS = set(klass.ESCAPE_FOLLOW_CHARS)
 566        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 567        klass._COMMENTS = {
 568            **dict(
 569                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 570                for comment in klass.COMMENTS
 571            ),
 572            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 573        }
 574        if klass.HINT_START in klass.KEYWORDS:
 575            klass._COMMENTS[klass.HINT_START] = "*/"
 576
 577        klass._KEYWORD_TRIE = new_trie(
 578            key.upper()
 579            for key in (
 580                *klass.KEYWORDS,
 581                *klass._COMMENTS,
 582                *klass._QUOTES,
 583                *klass._FORMAT_STRINGS,
 584            )
 585            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 586        )
 587
 588        if USE_RS_TOKENIZER:
 589            settings = RsTokenizerSettings(
 590                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 591                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 592                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 593                numeric_literals=klass.NUMERIC_LITERALS,
 594                identifiers=klass._IDENTIFIERS,
 595                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 596                string_escapes=klass._STRING_ESCAPES,
 597                byte_string_escapes=klass._BYTE_STRING_ESCAPES,
 598                quotes=klass._QUOTES,
 599                format_strings={
 600                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 601                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 602                },
 603                has_bit_strings=bool(klass.BIT_STRINGS),
 604                has_hex_strings=bool(klass.HEX_STRINGS),
 605                comments=klass._COMMENTS,
 606                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 607                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 608                command_prefix_tokens={
 609                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 610                },
 611                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 612                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 613                nested_comments=klass.NESTED_COMMENTS,
 614                hint_start=klass.HINT_START,
 615                tokens_preceding_hint={
 616                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 617                },
 618                escape_follow_chars=klass._ESCAPE_FOLLOW_CHARS,
 619            )
 620            token_types = RsTokenTypeSettings(
 621                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 622                byte_string=_TOKEN_TYPE_TO_INDEX[TokenType.BYTE_STRING],
 623                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 624                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 625                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 626                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 627                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 628                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 629                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 630                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 631                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 632                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 633                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 634                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 635                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 636            )
 637            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 638        else:
 639            klass._RS_TOKENIZER = None
 640
 641        return klass
 642
 643
 644class Tokenizer(metaclass=_Tokenizer):
 645    SINGLE_TOKENS = {
 646        "(": TokenType.L_PAREN,
 647        ")": TokenType.R_PAREN,
 648        "[": TokenType.L_BRACKET,
 649        "]": TokenType.R_BRACKET,
 650        "{": TokenType.L_BRACE,
 651        "}": TokenType.R_BRACE,
 652        "&": TokenType.AMP,
 653        "^": TokenType.CARET,
 654        ":": TokenType.COLON,
 655        ",": TokenType.COMMA,
 656        ".": TokenType.DOT,
 657        "-": TokenType.DASH,
 658        "=": TokenType.EQ,
 659        ">": TokenType.GT,
 660        "<": TokenType.LT,
 661        "%": TokenType.MOD,
 662        "!": TokenType.NOT,
 663        "|": TokenType.PIPE,
 664        "+": TokenType.PLUS,
 665        ";": TokenType.SEMICOLON,
 666        "/": TokenType.SLASH,
 667        "\\": TokenType.BACKSLASH,
 668        "*": TokenType.STAR,
 669        "~": TokenType.TILDE,
 670        "?": TokenType.PLACEHOLDER,
 671        "@": TokenType.PARAMETER,
 672        "#": TokenType.HASH,
 673        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 674        "'": TokenType.UNKNOWN,
 675        "`": TokenType.UNKNOWN,
 676        '"': TokenType.UNKNOWN,
 677    }
 678
 679    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 680    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 681    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 682    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 683    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 684    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 685    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 686    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 687    STRING_ESCAPES = ["'"]
 688    BYTE_STRING_ESCAPES: t.List[str] = []
 689    VAR_SINGLE_TOKENS: t.Set[str] = set()
 690    ESCAPE_FOLLOW_CHARS: t.List[str] = []
 691
 692    # The strings in this list can always be used as escapes, regardless of the surrounding
 693    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 694    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 695    IDENTIFIER_ESCAPES: t.List[str] = []
 696
 697    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 698    HEREDOC_TAG_IS_IDENTIFIER = False
 699
 700    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 701    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 702
 703    # Whether string escape characters function as such when placed within raw strings
 704    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 705
 706    NESTED_COMMENTS = True
 707
 708    HINT_START = "/*+"
 709
 710    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 711
 712    # Autofilled
 713    _COMMENTS: t.Dict[str, str] = {}
 714    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 715    _IDENTIFIERS: t.Dict[str, str] = {}
 716    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 717    _QUOTES: t.Dict[str, str] = {}
 718    _STRING_ESCAPES: t.Set[str] = set()
 719    _BYTE_STRING_ESCAPES: t.Set[str] = set()
 720    _KEYWORD_TRIE: t.Dict = {}
 721    _RS_TOKENIZER: t.Optional[t.Any] = None
 722    _ESCAPE_FOLLOW_CHARS: t.Set[str] = set()
 723
 724    KEYWORDS: t.Dict[str, TokenType] = {
 725        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 726        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 727        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 728        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 729        HINT_START: TokenType.HINT,
 730        "&<": TokenType.AMP_LT,
 731        "&>": TokenType.AMP_GT,
 732        "==": TokenType.EQ,
 733        "::": TokenType.DCOLON,
 734        "?::": TokenType.QDCOLON,
 735        "||": TokenType.DPIPE,
 736        "|>": TokenType.PIPE_GT,
 737        ">=": TokenType.GTE,
 738        "<=": TokenType.LTE,
 739        "<>": TokenType.NEQ,
 740        "!=": TokenType.NEQ,
 741        ":=": TokenType.COLON_EQ,
 742        "<=>": TokenType.NULLSAFE_EQ,
 743        "->": TokenType.ARROW,
 744        "->>": TokenType.DARROW,
 745        "=>": TokenType.FARROW,
 746        "#>": TokenType.HASH_ARROW,
 747        "#>>": TokenType.DHASH_ARROW,
 748        "<->": TokenType.LR_ARROW,
 749        "&&": TokenType.DAMP,
 750        "??": TokenType.DQMARK,
 751        "~~~": TokenType.GLOB,
 752        "~~": TokenType.LIKE,
 753        "~~*": TokenType.ILIKE,
 754        "~*": TokenType.IRLIKE,
 755        "-|-": TokenType.ADJACENT,
 756        "ALL": TokenType.ALL,
 757        "AND": TokenType.AND,
 758        "ANTI": TokenType.ANTI,
 759        "ANY": TokenType.ANY,
 760        "ASC": TokenType.ASC,
 761        "AS": TokenType.ALIAS,
 762        "ASOF": TokenType.ASOF,
 763        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 764        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 765        "BEGIN": TokenType.BEGIN,
 766        "BETWEEN": TokenType.BETWEEN,
 767        "CACHE": TokenType.CACHE,
 768        "UNCACHE": TokenType.UNCACHE,
 769        "CASE": TokenType.CASE,
 770        "CHARACTER SET": TokenType.CHARACTER_SET,
 771        "CLUSTER BY": TokenType.CLUSTER_BY,
 772        "COLLATE": TokenType.COLLATE,
 773        "COLUMN": TokenType.COLUMN,
 774        "COMMIT": TokenType.COMMIT,
 775        "CONNECT BY": TokenType.CONNECT_BY,
 776        "CONSTRAINT": TokenType.CONSTRAINT,
 777        "COPY": TokenType.COPY,
 778        "CREATE": TokenType.CREATE,
 779        "CROSS": TokenType.CROSS,
 780        "CUBE": TokenType.CUBE,
 781        "CURRENT_DATE": TokenType.CURRENT_DATE,
 782        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 783        "CURRENT_TIME": TokenType.CURRENT_TIME,
 784        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 785        "CURRENT_USER": TokenType.CURRENT_USER,
 786        "CURRENT_CATALOG": TokenType.CURRENT_CATALOG,
 787        "DATABASE": TokenType.DATABASE,
 788        "DEFAULT": TokenType.DEFAULT,
 789        "DELETE": TokenType.DELETE,
 790        "DESC": TokenType.DESC,
 791        "DESCRIBE": TokenType.DESCRIBE,
 792        "DISTINCT": TokenType.DISTINCT,
 793        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 794        "DIV": TokenType.DIV,
 795        "DROP": TokenType.DROP,
 796        "ELSE": TokenType.ELSE,
 797        "END": TokenType.END,
 798        "ENUM": TokenType.ENUM,
 799        "ESCAPE": TokenType.ESCAPE,
 800        "EXCEPT": TokenType.EXCEPT,
 801        "EXECUTE": TokenType.EXECUTE,
 802        "EXISTS": TokenType.EXISTS,
 803        "FALSE": TokenType.FALSE,
 804        "FETCH": TokenType.FETCH,
 805        "FILTER": TokenType.FILTER,
 806        "FILE": TokenType.FILE,
 807        "FIRST": TokenType.FIRST,
 808        "FULL": TokenType.FULL,
 809        "FUNCTION": TokenType.FUNCTION,
 810        "FOR": TokenType.FOR,
 811        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 812        "FORMAT": TokenType.FORMAT,
 813        "FROM": TokenType.FROM,
 814        "GEOGRAPHY": TokenType.GEOGRAPHY,
 815        "GEOMETRY": TokenType.GEOMETRY,
 816        "GLOB": TokenType.GLOB,
 817        "GROUP BY": TokenType.GROUP_BY,
 818        "GROUPING SETS": TokenType.GROUPING_SETS,
 819        "HAVING": TokenType.HAVING,
 820        "ILIKE": TokenType.ILIKE,
 821        "IN": TokenType.IN,
 822        "INDEX": TokenType.INDEX,
 823        "INET": TokenType.INET,
 824        "INNER": TokenType.INNER,
 825        "INSERT": TokenType.INSERT,
 826        "INTERVAL": TokenType.INTERVAL,
 827        "INTERSECT": TokenType.INTERSECT,
 828        "INTO": TokenType.INTO,
 829        "IS": TokenType.IS,
 830        "ISNULL": TokenType.ISNULL,
 831        "JOIN": TokenType.JOIN,
 832        "KEEP": TokenType.KEEP,
 833        "KILL": TokenType.KILL,
 834        "LATERAL": TokenType.LATERAL,
 835        "LEFT": TokenType.LEFT,
 836        "LIKE": TokenType.LIKE,
 837        "LIMIT": TokenType.LIMIT,
 838        "LOAD": TokenType.LOAD,
 839        "LOCALTIME": TokenType.LOCALTIME,
 840        "LOCALTIMESTAMP": TokenType.LOCALTIMESTAMP,
 841        "LOCK": TokenType.LOCK,
 842        "MERGE": TokenType.MERGE,
 843        "NAMESPACE": TokenType.NAMESPACE,
 844        "NATURAL": TokenType.NATURAL,
 845        "NEXT": TokenType.NEXT,
 846        "NOT": TokenType.NOT,
 847        "NOTNULL": TokenType.NOTNULL,
 848        "NULL": TokenType.NULL,
 849        "OBJECT": TokenType.OBJECT,
 850        "OFFSET": TokenType.OFFSET,
 851        "ON": TokenType.ON,
 852        "OR": TokenType.OR,
 853        "XOR": TokenType.XOR,
 854        "ORDER BY": TokenType.ORDER_BY,
 855        "ORDINALITY": TokenType.ORDINALITY,
 856        "OUT": TokenType.OUT,
 857        "OUTER": TokenType.OUTER,
 858        "OVER": TokenType.OVER,
 859        "OVERLAPS": TokenType.OVERLAPS,
 860        "OVERWRITE": TokenType.OVERWRITE,
 861        "PARTITION": TokenType.PARTITION,
 862        "PARTITION BY": TokenType.PARTITION_BY,
 863        "PARTITIONED BY": TokenType.PARTITION_BY,
 864        "PARTITIONED_BY": TokenType.PARTITION_BY,
 865        "PERCENT": TokenType.PERCENT,
 866        "PIVOT": TokenType.PIVOT,
 867        "PRAGMA": TokenType.PRAGMA,
 868        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 869        "PROCEDURE": TokenType.PROCEDURE,
 870        "OPERATOR": TokenType.OPERATOR,
 871        "QUALIFY": TokenType.QUALIFY,
 872        "RANGE": TokenType.RANGE,
 873        "RECURSIVE": TokenType.RECURSIVE,
 874        "REGEXP": TokenType.RLIKE,
 875        "RENAME": TokenType.RENAME,
 876        "REPLACE": TokenType.REPLACE,
 877        "RETURNING": TokenType.RETURNING,
 878        "REFERENCES": TokenType.REFERENCES,
 879        "RIGHT": TokenType.RIGHT,
 880        "RLIKE": TokenType.RLIKE,
 881        "ROLLBACK": TokenType.ROLLBACK,
 882        "ROLLUP": TokenType.ROLLUP,
 883        "ROW": TokenType.ROW,
 884        "ROWS": TokenType.ROWS,
 885        "SCHEMA": TokenType.SCHEMA,
 886        "SELECT": TokenType.SELECT,
 887        "SEMI": TokenType.SEMI,
 888        "SESSION": TokenType.SESSION,
 889        "SESSION_USER": TokenType.SESSION_USER,
 890        "SET": TokenType.SET,
 891        "SETTINGS": TokenType.SETTINGS,
 892        "SHOW": TokenType.SHOW,
 893        "SIMILAR TO": TokenType.SIMILAR_TO,
 894        "SOME": TokenType.SOME,
 895        "SORT BY": TokenType.SORT_BY,
 896        "START WITH": TokenType.START_WITH,
 897        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 898        "TABLE": TokenType.TABLE,
 899        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 900        "TEMP": TokenType.TEMPORARY,
 901        "TEMPORARY": TokenType.TEMPORARY,
 902        "THEN": TokenType.THEN,
 903        "TRUE": TokenType.TRUE,
 904        "TRUNCATE": TokenType.TRUNCATE,
 905        "UNION": TokenType.UNION,
 906        "UNKNOWN": TokenType.UNKNOWN,
 907        "UNNEST": TokenType.UNNEST,
 908        "UNPIVOT": TokenType.UNPIVOT,
 909        "UPDATE": TokenType.UPDATE,
 910        "USE": TokenType.USE,
 911        "USING": TokenType.USING,
 912        "UUID": TokenType.UUID,
 913        "VALUES": TokenType.VALUES,
 914        "VIEW": TokenType.VIEW,
 915        "VOLATILE": TokenType.VOLATILE,
 916        "WHEN": TokenType.WHEN,
 917        "WHERE": TokenType.WHERE,
 918        "WINDOW": TokenType.WINDOW,
 919        "WITH": TokenType.WITH,
 920        "APPLY": TokenType.APPLY,
 921        "ARRAY": TokenType.ARRAY,
 922        "BIT": TokenType.BIT,
 923        "BOOL": TokenType.BOOLEAN,
 924        "BOOLEAN": TokenType.BOOLEAN,
 925        "BYTE": TokenType.TINYINT,
 926        "MEDIUMINT": TokenType.MEDIUMINT,
 927        "INT1": TokenType.TINYINT,
 928        "TINYINT": TokenType.TINYINT,
 929        "INT16": TokenType.SMALLINT,
 930        "SHORT": TokenType.SMALLINT,
 931        "SMALLINT": TokenType.SMALLINT,
 932        "HUGEINT": TokenType.INT128,
 933        "UHUGEINT": TokenType.UINT128,
 934        "INT2": TokenType.SMALLINT,
 935        "INTEGER": TokenType.INT,
 936        "INT": TokenType.INT,
 937        "INT4": TokenType.INT,
 938        "INT32": TokenType.INT,
 939        "INT64": TokenType.BIGINT,
 940        "INT128": TokenType.INT128,
 941        "INT256": TokenType.INT256,
 942        "LONG": TokenType.BIGINT,
 943        "BIGINT": TokenType.BIGINT,
 944        "INT8": TokenType.TINYINT,
 945        "UINT": TokenType.UINT,
 946        "UINT128": TokenType.UINT128,
 947        "UINT256": TokenType.UINT256,
 948        "DEC": TokenType.DECIMAL,
 949        "DECIMAL": TokenType.DECIMAL,
 950        "DECIMAL32": TokenType.DECIMAL32,
 951        "DECIMAL64": TokenType.DECIMAL64,
 952        "DECIMAL128": TokenType.DECIMAL128,
 953        "DECIMAL256": TokenType.DECIMAL256,
 954        "DECFLOAT": TokenType.DECFLOAT,
 955        "BIGDECIMAL": TokenType.BIGDECIMAL,
 956        "BIGNUMERIC": TokenType.BIGDECIMAL,
 957        "BIGNUM": TokenType.BIGNUM,
 958        "LIST": TokenType.LIST,
 959        "MAP": TokenType.MAP,
 960        "NULLABLE": TokenType.NULLABLE,
 961        "NUMBER": TokenType.DECIMAL,
 962        "NUMERIC": TokenType.DECIMAL,
 963        "FIXED": TokenType.DECIMAL,
 964        "REAL": TokenType.FLOAT,
 965        "FLOAT": TokenType.FLOAT,
 966        "FLOAT4": TokenType.FLOAT,
 967        "FLOAT8": TokenType.DOUBLE,
 968        "DOUBLE": TokenType.DOUBLE,
 969        "DOUBLE PRECISION": TokenType.DOUBLE,
 970        "JSON": TokenType.JSON,
 971        "JSONB": TokenType.JSONB,
 972        "CHAR": TokenType.CHAR,
 973        "CHARACTER": TokenType.CHAR,
 974        "CHAR VARYING": TokenType.VARCHAR,
 975        "CHARACTER VARYING": TokenType.VARCHAR,
 976        "NCHAR": TokenType.NCHAR,
 977        "VARCHAR": TokenType.VARCHAR,
 978        "VARCHAR2": TokenType.VARCHAR,
 979        "NVARCHAR": TokenType.NVARCHAR,
 980        "NVARCHAR2": TokenType.NVARCHAR,
 981        "BPCHAR": TokenType.BPCHAR,
 982        "STR": TokenType.TEXT,
 983        "STRING": TokenType.TEXT,
 984        "TEXT": TokenType.TEXT,
 985        "LONGTEXT": TokenType.LONGTEXT,
 986        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 987        "TINYTEXT": TokenType.TINYTEXT,
 988        "CLOB": TokenType.TEXT,
 989        "LONGVARCHAR": TokenType.TEXT,
 990        "BINARY": TokenType.BINARY,
 991        "BLOB": TokenType.VARBINARY,
 992        "LONGBLOB": TokenType.LONGBLOB,
 993        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 994        "TINYBLOB": TokenType.TINYBLOB,
 995        "BYTEA": TokenType.VARBINARY,
 996        "VARBINARY": TokenType.VARBINARY,
 997        "TIME": TokenType.TIME,
 998        "TIMETZ": TokenType.TIMETZ,
 999        "TIME_NS": TokenType.TIME_NS,
1000        "TIMESTAMP": TokenType.TIMESTAMP,
1001        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
1002        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
1003        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
1004        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
1005        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
1006        "DATE": TokenType.DATE,
1007        "DATETIME": TokenType.DATETIME,
1008        "INT4RANGE": TokenType.INT4RANGE,
1009        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
1010        "INT8RANGE": TokenType.INT8RANGE,
1011        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
1012        "NUMRANGE": TokenType.NUMRANGE,
1013        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
1014        "TSRANGE": TokenType.TSRANGE,
1015        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
1016        "TSTZRANGE": TokenType.TSTZRANGE,
1017        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
1018        "DATERANGE": TokenType.DATERANGE,
1019        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
1020        "UNIQUE": TokenType.UNIQUE,
1021        "VECTOR": TokenType.VECTOR,
1022        "STRUCT": TokenType.STRUCT,
1023        "SEQUENCE": TokenType.SEQUENCE,
1024        "VARIANT": TokenType.VARIANT,
1025        "ALTER": TokenType.ALTER,
1026        "ANALYZE": TokenType.ANALYZE,
1027        "CALL": TokenType.COMMAND,
1028        "COMMENT": TokenType.COMMENT,
1029        "EXPLAIN": TokenType.COMMAND,
1030        "GRANT": TokenType.GRANT,
1031        "REVOKE": TokenType.REVOKE,
1032        "OPTIMIZE": TokenType.COMMAND,
1033        "PREPARE": TokenType.COMMAND,
1034        "VACUUM": TokenType.COMMAND,
1035        "USER-DEFINED": TokenType.USERDEFINED,
1036        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
1037        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
1038    }
1039
1040    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
1041        " ": TokenType.SPACE,
1042        "\t": TokenType.SPACE,
1043        "\n": TokenType.BREAK,
1044        "\r": TokenType.BREAK,
1045    }
1046
1047    COMMANDS = {
1048        TokenType.COMMAND,
1049        TokenType.EXECUTE,
1050        TokenType.FETCH,
1051        TokenType.SHOW,
1052        TokenType.RENAME,
1053    }
1054
1055    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
1056
1057    # Handle numeric literals like in hive (3L = BIGINT)
1058    NUMERIC_LITERALS: t.Dict[str, str] = {}
1059
1060    COMMENTS = ["--", ("/*", "*/")]
1061
1062    __slots__ = (
1063        "sql",
1064        "size",
1065        "tokens",
1066        "dialect",
1067        "use_rs_tokenizer",
1068        "_start",
1069        "_current",
1070        "_line",
1071        "_col",
1072        "_comments",
1073        "_char",
1074        "_end",
1075        "_peek",
1076        "_prev_token_line",
1077        "_rs_dialect_settings",
1078    )
1079
1080    def __init__(
1081        self,
1082        dialect: DialectType = None,
1083        use_rs_tokenizer: t.Optional[bool] = None,
1084        **opts: t.Any,
1085    ) -> None:
1086        from sqlglot.dialects import Dialect
1087
1088        self.dialect = Dialect.get_or_raise(dialect)
1089
1090        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1091        self.use_rs_tokenizer = (
1092            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1093        )
1094
1095        if self.use_rs_tokenizer:
1096            self._rs_dialect_settings = RsTokenizerDialectSettings(
1097                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1098                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1099                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1100            )
1101
1102        self.reset()
1103
1104    def reset(self) -> None:
1105        self.sql = ""
1106        self.size = 0
1107        self.tokens: t.List[Token] = []
1108        self._start = 0
1109        self._current = 0
1110        self._line = 1
1111        self._col = 0
1112        self._comments: t.List[str] = []
1113
1114        self._char = ""
1115        self._end = False
1116        self._peek = ""
1117        self._prev_token_line = -1
1118
1119    def tokenize(self, sql: str) -> t.List[Token]:
1120        """Returns a list of tokens corresponding to the SQL string `sql`."""
1121        if self.use_rs_tokenizer:
1122            return self.tokenize_rs(sql)
1123
1124        self.reset()
1125        self.sql = sql
1126        self.size = len(sql)
1127
1128        try:
1129            self._scan()
1130        except Exception as e:
1131            start = max(self._current - 50, 0)
1132            end = min(self._current + 50, self.size - 1)
1133            context = self.sql[start:end]
1134            raise TokenError(f"Error tokenizing '{context}'") from e
1135
1136        return self.tokens
1137
1138    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1139        while self.size and not self._end:
1140            current = self._current
1141
1142            # Skip spaces here rather than iteratively calling advance() for performance reasons
1143            while current < self.size:
1144                char = self.sql[current]
1145
1146                if char.isspace() and (char == " " or char == "\t"):
1147                    current += 1
1148                else:
1149                    break
1150
1151            offset = current - self._current if current > self._current else 1
1152
1153            self._start = current
1154            self._advance(offset)
1155
1156            if not self._char.isspace():
1157                if self._char.isdigit():
1158                    self._scan_number()
1159                elif self._char in self._IDENTIFIERS:
1160                    self._scan_identifier(self._IDENTIFIERS[self._char])
1161                else:
1162                    self._scan_keywords()
1163
1164            if until and until():
1165                break
1166
1167        if self.tokens and self._comments:
1168            self.tokens[-1].comments.extend(self._comments)
1169
1170    def _chars(self, size: int) -> str:
1171        if size == 1:
1172            return self._char
1173
1174        start = self._current - 1
1175        end = start + size
1176
1177        return self.sql[start:end] if end <= self.size else ""
1178
1179    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1180        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1181            # Ensures we don't count an extra line if we get a \r\n line break sequence
1182            if not (self._char == "\r" and self._peek == "\n"):
1183                self._col = i
1184                self._line += 1
1185        else:
1186            self._col += i
1187
1188        self._current += i
1189        self._end = self._current >= self.size
1190        self._char = self.sql[self._current - 1]
1191        self._peek = "" if self._end else self.sql[self._current]
1192
1193        if alnum and self._char.isalnum():
1194            # Here we use local variables instead of attributes for better performance
1195            _col = self._col
1196            _current = self._current
1197            _end = self._end
1198            _peek = self._peek
1199
1200            while _peek.isalnum():
1201                _col += 1
1202                _current += 1
1203                _end = _current >= self.size
1204                _peek = "" if _end else self.sql[_current]
1205
1206            self._col = _col
1207            self._current = _current
1208            self._end = _end
1209            self._peek = _peek
1210            self._char = self.sql[_current - 1]
1211
1212    @property
1213    def _text(self) -> str:
1214        return self.sql[self._start : self._current]
1215
1216    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1217        self._prev_token_line = self._line
1218
1219        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1220            self.tokens[-1].comments.extend(self._comments)
1221            self._comments = []
1222
1223        self.tokens.append(
1224            Token(
1225                token_type,
1226                text=self._text if text is None else text,
1227                line=self._line,
1228                col=self._col,
1229                start=self._start,
1230                end=self._current - 1,
1231                comments=self._comments,
1232            )
1233        )
1234        self._comments = []
1235
1236        # If we have either a semicolon or a begin token before the command's token, we'll parse
1237        # whatever follows the command's token as a string
1238        if (
1239            token_type in self.COMMANDS
1240            and self._peek != ";"
1241            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1242        ):
1243            start = self._current
1244            tokens = len(self.tokens)
1245            self._scan(lambda: self._peek == ";")
1246            self.tokens = self.tokens[:tokens]
1247            text = self.sql[start : self._current].strip()
1248            if text:
1249                self._add(TokenType.STRING, text)
1250
1251    def _scan_keywords(self) -> None:
1252        size = 0
1253        word = None
1254        chars = self._text
1255        char = chars
1256        prev_space = False
1257        skip = False
1258        trie = self._KEYWORD_TRIE
1259        single_token = char in self.SINGLE_TOKENS
1260
1261        while chars:
1262            if skip:
1263                result = TrieResult.PREFIX
1264            else:
1265                result, trie = in_trie(trie, char.upper())
1266
1267            if result == TrieResult.FAILED:
1268                break
1269            if result == TrieResult.EXISTS:
1270                word = chars
1271
1272            end = self._current + size
1273            size += 1
1274
1275            if end < self.size:
1276                char = self.sql[end]
1277                single_token = single_token or char in self.SINGLE_TOKENS
1278                is_space = char.isspace()
1279
1280                if not is_space or not prev_space:
1281                    if is_space:
1282                        char = " "
1283                    chars += char
1284                    prev_space = is_space
1285                    skip = False
1286                else:
1287                    skip = True
1288            else:
1289                char = ""
1290                break
1291
1292        if word:
1293            if self._scan_string(word):
1294                return
1295            if self._scan_comment(word):
1296                return
1297            if prev_space or single_token or not char:
1298                self._advance(size - 1)
1299                word = word.upper()
1300                self._add(self.KEYWORDS[word], text=word)
1301                return
1302
1303        if self._char in self.SINGLE_TOKENS:
1304            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1305            return
1306
1307        self._scan_var()
1308
1309    def _scan_comment(self, comment_start: str) -> bool:
1310        if comment_start not in self._COMMENTS:
1311            return False
1312
1313        comment_start_line = self._line
1314        comment_start_size = len(comment_start)
1315        comment_end = self._COMMENTS[comment_start]
1316
1317        if comment_end:
1318            # Skip the comment's start delimiter
1319            self._advance(comment_start_size)
1320
1321            comment_count = 1
1322            comment_end_size = len(comment_end)
1323
1324            while not self._end:
1325                if self._chars(comment_end_size) == comment_end:
1326                    comment_count -= 1
1327                    if not comment_count:
1328                        break
1329
1330                self._advance(alnum=True)
1331
1332                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1333                if (
1334                    self.NESTED_COMMENTS
1335                    and not self._end
1336                    and self._chars(comment_end_size) == comment_start
1337                ):
1338                    self._advance(comment_start_size)
1339                    comment_count += 1
1340
1341            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1342            self._advance(comment_end_size - 1)
1343        else:
1344            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1345                self._advance(alnum=True)
1346            self._comments.append(self._text[comment_start_size:])
1347
1348        if (
1349            comment_start == self.HINT_START
1350            and self.tokens
1351            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1352        ):
1353            self._add(TokenType.HINT)
1354
1355        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1356        # Multiple consecutive comments are preserved by appending them to the current comments list.
1357        if comment_start_line == self._prev_token_line:
1358            self.tokens[-1].comments.extend(self._comments)
1359            self._comments = []
1360            self._prev_token_line = self._line
1361
1362        return True
1363
1364    def _scan_number(self) -> None:
1365        if self._char == "0":
1366            peek = self._peek.upper()
1367            if peek == "B":
1368                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1369            elif peek == "X":
1370                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1371
1372        decimal = False
1373        scientific = 0
1374
1375        while True:
1376            if self._peek.isdigit():
1377                self._advance()
1378            elif self._peek == "." and not decimal:
1379                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1380                    return self._add(TokenType.NUMBER)
1381                decimal = True
1382                self._advance()
1383            elif self._peek in ("-", "+") and scientific == 1:
1384                # Only consume +/- if followed by a digit
1385                if self._current + 1 < self.size and self.sql[self._current + 1].isdigit():
1386                    scientific += 1
1387                    self._advance()
1388                else:
1389                    return self._add(TokenType.NUMBER)
1390            elif self._peek.upper() == "E" and not scientific:
1391                scientific += 1
1392                self._advance()
1393            elif self._peek == "_" and self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED:
1394                self._advance()
1395            elif self._peek.isidentifier():
1396                number_text = self._text
1397                literal = ""
1398
1399                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1400                    literal += self._peek
1401                    self._advance()
1402
1403                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1404
1405                if token_type:
1406                    self._add(TokenType.NUMBER, number_text)
1407                    self._add(TokenType.DCOLON, "::")
1408                    return self._add(token_type, literal)
1409                elif self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1410                    return self._add(TokenType.VAR)
1411
1412                self._advance(-len(literal))
1413                return self._add(TokenType.NUMBER, number_text)
1414            else:
1415                return self._add(TokenType.NUMBER)
1416
1417    def _scan_bits(self) -> None:
1418        self._advance()
1419        value = self._extract_value()
1420        try:
1421            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1422            int(value, 2)
1423            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1424        except ValueError:
1425            self._add(TokenType.IDENTIFIER)
1426
1427    def _scan_hex(self) -> None:
1428        self._advance()
1429        value = self._extract_value()
1430        try:
1431            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1432            int(value, 16)
1433            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1434        except ValueError:
1435            self._add(TokenType.IDENTIFIER)
1436
1437    def _extract_value(self) -> str:
1438        while True:
1439            char = self._peek.strip()
1440            if char and char not in self.SINGLE_TOKENS:
1441                self._advance(alnum=True)
1442            else:
1443                break
1444
1445        return self._text
1446
1447    def _scan_string(self, start: str) -> bool:
1448        base = None
1449        token_type = TokenType.STRING
1450
1451        if start in self._QUOTES:
1452            end = self._QUOTES[start]
1453        elif start in self._FORMAT_STRINGS:
1454            end, token_type = self._FORMAT_STRINGS[start]
1455
1456            if token_type == TokenType.HEX_STRING:
1457                base = 16
1458            elif token_type == TokenType.BIT_STRING:
1459                base = 2
1460            elif token_type == TokenType.HEREDOC_STRING:
1461                self._advance()
1462
1463                if self._char == end:
1464                    tag = ""
1465                else:
1466                    tag = self._extract_string(
1467                        end,
1468                        raw_string=True,
1469                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1470                    )
1471
1472                if (
1473                    tag
1474                    and self.HEREDOC_TAG_IS_IDENTIFIER
1475                    and (self._end or tag.isdigit() or any(c.isspace() for c in tag))
1476                ):
1477                    if not self._end:
1478                        self._advance(-1)
1479
1480                    self._advance(-len(tag))
1481                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1482                    return True
1483
1484                end = f"{start}{tag}{end}"
1485        else:
1486            return False
1487
1488        self._advance(len(start))
1489        text = self._extract_string(
1490            end,
1491            escapes=(
1492                self._BYTE_STRING_ESCAPES
1493                if token_type == TokenType.BYTE_STRING
1494                else self._STRING_ESCAPES
1495            ),
1496            raw_string=token_type == TokenType.RAW_STRING,
1497        )
1498
1499        if base and text:
1500            try:
1501                int(text, base)
1502            except Exception:
1503                raise TokenError(
1504                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1505                )
1506
1507        self._add(token_type, text)
1508        return True
1509
1510    def _scan_identifier(self, identifier_end: str) -> None:
1511        self._advance()
1512        text = self._extract_string(
1513            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1514        )
1515        self._add(TokenType.IDENTIFIER, text)
1516
1517    def _scan_var(self) -> None:
1518        while True:
1519            char = self._peek.strip()
1520            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1521                self._advance(alnum=True)
1522            else:
1523                break
1524
1525        self._add(
1526            TokenType.VAR
1527            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1528            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1529        )
1530
1531    def _extract_string(
1532        self,
1533        delimiter: str,
1534        escapes: t.Optional[t.Set[str]] = None,
1535        raw_string: bool = False,
1536        raise_unmatched: bool = True,
1537    ) -> str:
1538        text = ""
1539        delim_size = len(delimiter)
1540        escapes = self._STRING_ESCAPES if escapes is None else escapes
1541
1542        while True:
1543            if (
1544                not raw_string
1545                and self.dialect.UNESCAPED_SEQUENCES
1546                and self._peek
1547                and self._char in escapes
1548            ):
1549                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1550                if unescaped_sequence:
1551                    self._advance(2)
1552                    text += unescaped_sequence
1553                    continue
1554
1555            is_valid_custom_escape = (
1556                self.ESCAPE_FOLLOW_CHARS
1557                and self._char == "\\"
1558                and self._peek not in self.ESCAPE_FOLLOW_CHARS
1559            )
1560
1561            if (
1562                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1563                and self._char in escapes
1564                and (self._peek == delimiter or self._peek in escapes or is_valid_custom_escape)
1565                and (self._char not in self._QUOTES or self._char == self._peek)
1566            ):
1567                if self._peek == delimiter:
1568                    text += self._peek
1569                elif is_valid_custom_escape and self._char != self._peek:
1570                    text += self._peek
1571                else:
1572                    text += self._char + self._peek
1573
1574                if self._current + 1 < self.size:
1575                    self._advance(2)
1576                else:
1577                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1578            else:
1579                if self._chars(delim_size) == delimiter:
1580                    if delim_size > 1:
1581                        self._advance(delim_size - 1)
1582                    break
1583
1584                if self._end:
1585                    if not raise_unmatched:
1586                        return text + self._char
1587
1588                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1589
1590                current = self._current - 1
1591                self._advance(alnum=True)
1592                text += self.sql[current : self._current - 1]
1593
1594        return text
1595
1596    def tokenize_rs(self, sql: str) -> t.List[Token]:
1597        if not self._RS_TOKENIZER:
1598            raise SqlglotError("Rust tokenizer is not available")
1599
1600        tokens, error_msg = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1601        for token in tokens:
1602            token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1603
1604        # Setting this here so partial token lists can be inspected even if there is a failure
1605        self.tokens = tokens
1606
1607        if error_msg is not None:
1608            raise TokenError(error_msg)
1609
1610        return tokens