sqlglot.tokens

   1from __future__ import annotations
   2
   3import typing as t
   4from enum import auto
   5
   6from sqlglot.errors import TokenError
   7from sqlglot.helper import AutoName
   8from sqlglot.trie import TrieResult, in_trie, new_trie
   9
  10
  11class TokenType(AutoName):
  12    L_PAREN = auto()
  13    R_PAREN = auto()
  14    L_BRACKET = auto()
  15    R_BRACKET = auto()
  16    L_BRACE = auto()
  17    R_BRACE = auto()
  18    COMMA = auto()
  19    DOT = auto()
  20    DASH = auto()
  21    PLUS = auto()
  22    COLON = auto()
  23    DCOLON = auto()
  24    DQMARK = auto()
  25    SEMICOLON = auto()
  26    STAR = auto()
  27    BACKSLASH = auto()
  28    SLASH = auto()
  29    LT = auto()
  30    LTE = auto()
  31    GT = auto()
  32    GTE = auto()
  33    NOT = auto()
  34    EQ = auto()
  35    NEQ = auto()
  36    NULLSAFE_EQ = auto()
  37    AND = auto()
  38    OR = auto()
  39    AMP = auto()
  40    DPIPE = auto()
  41    PIPE = auto()
  42    CARET = auto()
  43    TILDA = auto()
  44    ARROW = auto()
  45    DARROW = auto()
  46    FARROW = auto()
  47    HASH = auto()
  48    HASH_ARROW = auto()
  49    DHASH_ARROW = auto()
  50    LR_ARROW = auto()
  51    DAT = auto()
  52    LT_AT = auto()
  53    AT_GT = auto()
  54    DOLLAR = auto()
  55    PARAMETER = auto()
  56    SESSION_PARAMETER = auto()
  57    DAMP = auto()
  58    XOR = auto()
  59
  60    BLOCK_START = auto()
  61    BLOCK_END = auto()
  62
  63    SPACE = auto()
  64    BREAK = auto()
  65
  66    STRING = auto()
  67    NUMBER = auto()
  68    IDENTIFIER = auto()
  69    DATABASE = auto()
  70    COLUMN = auto()
  71    COLUMN_DEF = auto()
  72    SCHEMA = auto()
  73    TABLE = auto()
  74    VAR = auto()
  75    BIT_STRING = auto()
  76    HEX_STRING = auto()
  77    BYTE_STRING = auto()
  78    NATIONAL_STRING = auto()
  79    RAW_STRING = auto()
  80
  81    # types
  82    BIT = auto()
  83    BOOLEAN = auto()
  84    TINYINT = auto()
  85    UTINYINT = auto()
  86    SMALLINT = auto()
  87    USMALLINT = auto()
  88    MEDIUMINT = auto()
  89    INT = auto()
  90    UINT = auto()
  91    BIGINT = auto()
  92    UBIGINT = auto()
  93    INT128 = auto()
  94    UINT128 = auto()
  95    INT256 = auto()
  96    UINT256 = auto()
  97    FLOAT = auto()
  98    DOUBLE = auto()
  99    DECIMAL = auto()
 100    BIGDECIMAL = auto()
 101    CHAR = auto()
 102    NCHAR = auto()
 103    VARCHAR = auto()
 104    NVARCHAR = auto()
 105    TEXT = auto()
 106    MEDIUMTEXT = auto()
 107    LONGTEXT = auto()
 108    MEDIUMBLOB = auto()
 109    LONGBLOB = auto()
 110    BINARY = auto()
 111    VARBINARY = auto()
 112    JSON = auto()
 113    JSONB = auto()
 114    TIME = auto()
 115    TIMETZ = auto()
 116    TIMESTAMP = auto()
 117    TIMESTAMPTZ = auto()
 118    TIMESTAMPLTZ = auto()
 119    DATETIME = auto()
 120    DATETIME64 = auto()
 121    DATE = auto()
 122    INT4RANGE = auto()
 123    INT4MULTIRANGE = auto()
 124    INT8RANGE = auto()
 125    INT8MULTIRANGE = auto()
 126    NUMRANGE = auto()
 127    NUMMULTIRANGE = auto()
 128    TSRANGE = auto()
 129    TSMULTIRANGE = auto()
 130    TSTZRANGE = auto()
 131    TSTZMULTIRANGE = auto()
 132    DATERANGE = auto()
 133    DATEMULTIRANGE = auto()
 134    UUID = auto()
 135    GEOGRAPHY = auto()
 136    NULLABLE = auto()
 137    GEOMETRY = auto()
 138    HLLSKETCH = auto()
 139    HSTORE = auto()
 140    SUPER = auto()
 141    SERIAL = auto()
 142    SMALLSERIAL = auto()
 143    BIGSERIAL = auto()
 144    XML = auto()
 145    YEAR = auto()
 146    UNIQUEIDENTIFIER = auto()
 147    USERDEFINED = auto()
 148    MONEY = auto()
 149    SMALLMONEY = auto()
 150    ROWVERSION = auto()
 151    IMAGE = auto()
 152    VARIANT = auto()
 153    OBJECT = auto()
 154    INET = auto()
 155    IPADDRESS = auto()
 156    IPPREFIX = auto()
 157    ENUM = auto()
 158    ENUM8 = auto()
 159    ENUM16 = auto()
 160    FIXEDSTRING = auto()
 161    LOWCARDINALITY = auto()
 162    NESTED = auto()
 163    UNKNOWN = auto()
 164
 165    # keywords
 166    ALIAS = auto()
 167    ALTER = auto()
 168    ALWAYS = auto()
 169    ALL = auto()
 170    ANTI = auto()
 171    ANY = auto()
 172    APPLY = auto()
 173    ARRAY = auto()
 174    ASC = auto()
 175    ASOF = auto()
 176    AUTO_INCREMENT = auto()
 177    BEGIN = auto()
 178    BETWEEN = auto()
 179    CACHE = auto()
 180    CASE = auto()
 181    CHARACTER_SET = auto()
 182    CLUSTER_BY = auto()
 183    COLLATE = auto()
 184    COMMAND = auto()
 185    COMMENT = auto()
 186    COMMIT = auto()
 187    CONNECT_BY = auto()
 188    CONSTRAINT = auto()
 189    CREATE = auto()
 190    CROSS = auto()
 191    CUBE = auto()
 192    CURRENT_DATE = auto()
 193    CURRENT_DATETIME = auto()
 194    CURRENT_TIME = auto()
 195    CURRENT_TIMESTAMP = auto()
 196    CURRENT_USER = auto()
 197    DEFAULT = auto()
 198    DELETE = auto()
 199    DESC = auto()
 200    DESCRIBE = auto()
 201    DICTIONARY = auto()
 202    DISTINCT = auto()
 203    DISTRIBUTE_BY = auto()
 204    DIV = auto()
 205    DROP = auto()
 206    ELSE = auto()
 207    END = auto()
 208    ESCAPE = auto()
 209    EXCEPT = auto()
 210    EXECUTE = auto()
 211    EXISTS = auto()
 212    FALSE = auto()
 213    FETCH = auto()
 214    FILTER = auto()
 215    FINAL = auto()
 216    FIRST = auto()
 217    FOR = auto()
 218    FORCE = auto()
 219    FOREIGN_KEY = auto()
 220    FORMAT = auto()
 221    FROM = auto()
 222    FULL = auto()
 223    FUNCTION = auto()
 224    GLOB = auto()
 225    GLOBAL = auto()
 226    GROUP_BY = auto()
 227    GROUPING_SETS = auto()
 228    HAVING = auto()
 229    HINT = auto()
 230    IGNORE = auto()
 231    ILIKE = auto()
 232    ILIKE_ANY = auto()
 233    IN = auto()
 234    INDEX = auto()
 235    INNER = auto()
 236    INSERT = auto()
 237    INTERSECT = auto()
 238    INTERVAL = auto()
 239    INTO = auto()
 240    INTRODUCER = auto()
 241    IRLIKE = auto()
 242    IS = auto()
 243    ISNULL = auto()
 244    JOIN = auto()
 245    JOIN_MARKER = auto()
 246    KEEP = auto()
 247    LANGUAGE = auto()
 248    LATERAL = auto()
 249    LEFT = auto()
 250    LIKE = auto()
 251    LIKE_ANY = auto()
 252    LIMIT = auto()
 253    LOAD = auto()
 254    LOCK = auto()
 255    MAP = auto()
 256    MATCH_RECOGNIZE = auto()
 257    MEMBER_OF = auto()
 258    MERGE = auto()
 259    MOD = auto()
 260    NATURAL = auto()
 261    NEXT = auto()
 262    NOTNULL = auto()
 263    NULL = auto()
 264    OBJECT_IDENTIFIER = auto()
 265    OFFSET = auto()
 266    ON = auto()
 267    ORDER_BY = auto()
 268    ORDERED = auto()
 269    ORDINALITY = auto()
 270    OUTER = auto()
 271    OVER = auto()
 272    OVERLAPS = auto()
 273    OVERWRITE = auto()
 274    PARTITION = auto()
 275    PARTITION_BY = auto()
 276    PERCENT = auto()
 277    PIVOT = auto()
 278    PLACEHOLDER = auto()
 279    PRAGMA = auto()
 280    PRIMARY_KEY = auto()
 281    PROCEDURE = auto()
 282    PROPERTIES = auto()
 283    PSEUDO_TYPE = auto()
 284    QUALIFY = auto()
 285    QUOTE = auto()
 286    RANGE = auto()
 287    RECURSIVE = auto()
 288    REPLACE = auto()
 289    RETURNING = auto()
 290    REFERENCES = auto()
 291    RIGHT = auto()
 292    RLIKE = auto()
 293    ROLLBACK = auto()
 294    ROLLUP = auto()
 295    ROW = auto()
 296    ROWS = auto()
 297    SELECT = auto()
 298    SEMI = auto()
 299    SEPARATOR = auto()
 300    SERDE_PROPERTIES = auto()
 301    SET = auto()
 302    SETTINGS = auto()
 303    SHOW = auto()
 304    SIMILAR_TO = auto()
 305    SOME = auto()
 306    SORT_BY = auto()
 307    START_WITH = auto()
 308    STRUCT = auto()
 309    TABLE_SAMPLE = auto()
 310    TEMPORARY = auto()
 311    TOP = auto()
 312    THEN = auto()
 313    TRUE = auto()
 314    UNCACHE = auto()
 315    UNION = auto()
 316    UNNEST = auto()
 317    UNPIVOT = auto()
 318    UPDATE = auto()
 319    USE = auto()
 320    USING = auto()
 321    VALUES = auto()
 322    VIEW = auto()
 323    VOLATILE = auto()
 324    WHEN = auto()
 325    WHERE = auto()
 326    WINDOW = auto()
 327    WITH = auto()
 328    UNIQUE = auto()
 329    VERSION_SNAPSHOT = auto()
 330    TIMESTAMP_SNAPSHOT = auto()
 331
 332
 333class Token:
 334    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 335
 336    @classmethod
 337    def number(cls, number: int) -> Token:
 338        """Returns a NUMBER token with `number` as its text."""
 339        return cls(TokenType.NUMBER, str(number))
 340
 341    @classmethod
 342    def string(cls, string: str) -> Token:
 343        """Returns a STRING token with `string` as its text."""
 344        return cls(TokenType.STRING, string)
 345
 346    @classmethod
 347    def identifier(cls, identifier: str) -> Token:
 348        """Returns an IDENTIFIER token with `identifier` as its text."""
 349        return cls(TokenType.IDENTIFIER, identifier)
 350
 351    @classmethod
 352    def var(cls, var: str) -> Token:
 353        """Returns an VAR token with `var` as its text."""
 354        return cls(TokenType.VAR, var)
 355
 356    def __init__(
 357        self,
 358        token_type: TokenType,
 359        text: str,
 360        line: int = 1,
 361        col: int = 1,
 362        start: int = 0,
 363        end: int = 0,
 364        comments: t.List[str] = [],
 365    ) -> None:
 366        """Token initializer.
 367
 368        Args:
 369            token_type: The TokenType Enum.
 370            text: The text of the token.
 371            line: The line that the token ends on.
 372            col: The column that the token ends on.
 373            start: The start index of the token.
 374            end: The ending index of the token.
 375            comments: The comments to attach to the token.
 376        """
 377        self.token_type = token_type
 378        self.text = text
 379        self.line = line
 380        self.col = col
 381        self.start = start
 382        self.end = end
 383        self.comments = comments
 384
 385    def __repr__(self) -> str:
 386        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 387        return f"<Token {attributes}>"
 388
 389
 390class _Tokenizer(type):
 391    def __new__(cls, clsname, bases, attrs):
 392        klass = super().__new__(cls, clsname, bases, attrs)
 393
 394        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 395            return dict(
 396                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 397            )
 398
 399        def _quotes_to_format(
 400            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 401        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 402            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 403
 404        klass._QUOTES = _convert_quotes(klass.QUOTES)
 405        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 406
 407        klass._FORMAT_STRINGS = {
 408            **{
 409                p + s: (e, TokenType.NATIONAL_STRING)
 410                for s, e in klass._QUOTES.items()
 411                for p in ("n", "N")
 412            },
 413            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 414            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 415            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 416            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 417        }
 418
 419        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 420        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 421        klass._COMMENTS = {
 422            **dict(
 423                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 424                for comment in klass.COMMENTS
 425            ),
 426            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 427        }
 428
 429        klass._KEYWORD_TRIE = new_trie(
 430            key.upper()
 431            for key in (
 432                *klass.KEYWORDS,
 433                *klass._COMMENTS,
 434                *klass._QUOTES,
 435                *klass._FORMAT_STRINGS,
 436            )
 437            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 438        )
 439
 440        return klass
 441
 442
 443class Tokenizer(metaclass=_Tokenizer):
 444    SINGLE_TOKENS = {
 445        "(": TokenType.L_PAREN,
 446        ")": TokenType.R_PAREN,
 447        "[": TokenType.L_BRACKET,
 448        "]": TokenType.R_BRACKET,
 449        "{": TokenType.L_BRACE,
 450        "}": TokenType.R_BRACE,
 451        "&": TokenType.AMP,
 452        "^": TokenType.CARET,
 453        ":": TokenType.COLON,
 454        ",": TokenType.COMMA,
 455        ".": TokenType.DOT,
 456        "-": TokenType.DASH,
 457        "=": TokenType.EQ,
 458        ">": TokenType.GT,
 459        "<": TokenType.LT,
 460        "%": TokenType.MOD,
 461        "!": TokenType.NOT,
 462        "|": TokenType.PIPE,
 463        "+": TokenType.PLUS,
 464        ";": TokenType.SEMICOLON,
 465        "/": TokenType.SLASH,
 466        "\\": TokenType.BACKSLASH,
 467        "*": TokenType.STAR,
 468        "~": TokenType.TILDA,
 469        "?": TokenType.PLACEHOLDER,
 470        "@": TokenType.PARAMETER,
 471        # used for breaking a var like x'y' but nothing else
 472        # the token type doesn't matter
 473        "'": TokenType.QUOTE,
 474        "`": TokenType.IDENTIFIER,
 475        '"': TokenType.IDENTIFIER,
 476        "#": TokenType.HASH,
 477    }
 478
 479    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 480    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 481    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 482    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 483    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 484    IDENTIFIER_ESCAPES = ['"']
 485    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 486    STRING_ESCAPES = ["'"]
 487    VAR_SINGLE_TOKENS: t.Set[str] = set()
 488
 489    # Autofilled
 490    IDENTIFIERS_CAN_START_WITH_DIGIT: bool = False
 491
 492    _COMMENTS: t.Dict[str, str] = {}
 493    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 494    _IDENTIFIERS: t.Dict[str, str] = {}
 495    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 496    _QUOTES: t.Dict[str, str] = {}
 497    _STRING_ESCAPES: t.Set[str] = set()
 498    _KEYWORD_TRIE: t.Dict = {}
 499
 500    KEYWORDS: t.Dict[str, TokenType] = {
 501        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 502        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 503        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 504        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 505        "/*+": TokenType.HINT,
 506        "==": TokenType.EQ,
 507        "::": TokenType.DCOLON,
 508        "||": TokenType.DPIPE,
 509        ">=": TokenType.GTE,
 510        "<=": TokenType.LTE,
 511        "<>": TokenType.NEQ,
 512        "!=": TokenType.NEQ,
 513        "<=>": TokenType.NULLSAFE_EQ,
 514        "->": TokenType.ARROW,
 515        "->>": TokenType.DARROW,
 516        "=>": TokenType.FARROW,
 517        "#>": TokenType.HASH_ARROW,
 518        "#>>": TokenType.DHASH_ARROW,
 519        "<->": TokenType.LR_ARROW,
 520        "&&": TokenType.DAMP,
 521        "??": TokenType.DQMARK,
 522        "ALL": TokenType.ALL,
 523        "ALWAYS": TokenType.ALWAYS,
 524        "AND": TokenType.AND,
 525        "ANTI": TokenType.ANTI,
 526        "ANY": TokenType.ANY,
 527        "ASC": TokenType.ASC,
 528        "AS": TokenType.ALIAS,
 529        "ASOF": TokenType.ASOF,
 530        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 531        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 532        "BEGIN": TokenType.BEGIN,
 533        "BETWEEN": TokenType.BETWEEN,
 534        "CACHE": TokenType.CACHE,
 535        "UNCACHE": TokenType.UNCACHE,
 536        "CASE": TokenType.CASE,
 537        "CHARACTER SET": TokenType.CHARACTER_SET,
 538        "CLUSTER BY": TokenType.CLUSTER_BY,
 539        "COLLATE": TokenType.COLLATE,
 540        "COLUMN": TokenType.COLUMN,
 541        "COMMIT": TokenType.COMMIT,
 542        "CONNECT BY": TokenType.CONNECT_BY,
 543        "CONSTRAINT": TokenType.CONSTRAINT,
 544        "CREATE": TokenType.CREATE,
 545        "CROSS": TokenType.CROSS,
 546        "CUBE": TokenType.CUBE,
 547        "CURRENT_DATE": TokenType.CURRENT_DATE,
 548        "CURRENT_TIME": TokenType.CURRENT_TIME,
 549        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 550        "CURRENT_USER": TokenType.CURRENT_USER,
 551        "DATABASE": TokenType.DATABASE,
 552        "DEFAULT": TokenType.DEFAULT,
 553        "DELETE": TokenType.DELETE,
 554        "DESC": TokenType.DESC,
 555        "DESCRIBE": TokenType.DESCRIBE,
 556        "DISTINCT": TokenType.DISTINCT,
 557        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 558        "DIV": TokenType.DIV,
 559        "DROP": TokenType.DROP,
 560        "ELSE": TokenType.ELSE,
 561        "END": TokenType.END,
 562        "ESCAPE": TokenType.ESCAPE,
 563        "EXCEPT": TokenType.EXCEPT,
 564        "EXECUTE": TokenType.EXECUTE,
 565        "EXISTS": TokenType.EXISTS,
 566        "FALSE": TokenType.FALSE,
 567        "FETCH": TokenType.FETCH,
 568        "FILTER": TokenType.FILTER,
 569        "FIRST": TokenType.FIRST,
 570        "FULL": TokenType.FULL,
 571        "FUNCTION": TokenType.FUNCTION,
 572        "FOR": TokenType.FOR,
 573        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 574        "FORMAT": TokenType.FORMAT,
 575        "FROM": TokenType.FROM,
 576        "GEOGRAPHY": TokenType.GEOGRAPHY,
 577        "GEOMETRY": TokenType.GEOMETRY,
 578        "GLOB": TokenType.GLOB,
 579        "GROUP BY": TokenType.GROUP_BY,
 580        "GROUPING SETS": TokenType.GROUPING_SETS,
 581        "HAVING": TokenType.HAVING,
 582        "ILIKE": TokenType.ILIKE,
 583        "IN": TokenType.IN,
 584        "INDEX": TokenType.INDEX,
 585        "INET": TokenType.INET,
 586        "INNER": TokenType.INNER,
 587        "INSERT": TokenType.INSERT,
 588        "INTERVAL": TokenType.INTERVAL,
 589        "INTERSECT": TokenType.INTERSECT,
 590        "INTO": TokenType.INTO,
 591        "IS": TokenType.IS,
 592        "ISNULL": TokenType.ISNULL,
 593        "JOIN": TokenType.JOIN,
 594        "KEEP": TokenType.KEEP,
 595        "LATERAL": TokenType.LATERAL,
 596        "LEFT": TokenType.LEFT,
 597        "LIKE": TokenType.LIKE,
 598        "LIMIT": TokenType.LIMIT,
 599        "LOAD": TokenType.LOAD,
 600        "LOCK": TokenType.LOCK,
 601        "MERGE": TokenType.MERGE,
 602        "NATURAL": TokenType.NATURAL,
 603        "NEXT": TokenType.NEXT,
 604        "NOT": TokenType.NOT,
 605        "NOTNULL": TokenType.NOTNULL,
 606        "NULL": TokenType.NULL,
 607        "OBJECT": TokenType.OBJECT,
 608        "OFFSET": TokenType.OFFSET,
 609        "ON": TokenType.ON,
 610        "OR": TokenType.OR,
 611        "XOR": TokenType.XOR,
 612        "ORDER BY": TokenType.ORDER_BY,
 613        "ORDINALITY": TokenType.ORDINALITY,
 614        "OUTER": TokenType.OUTER,
 615        "OVER": TokenType.OVER,
 616        "OVERLAPS": TokenType.OVERLAPS,
 617        "OVERWRITE": TokenType.OVERWRITE,
 618        "PARTITION": TokenType.PARTITION,
 619        "PARTITION BY": TokenType.PARTITION_BY,
 620        "PARTITIONED BY": TokenType.PARTITION_BY,
 621        "PARTITIONED_BY": TokenType.PARTITION_BY,
 622        "PERCENT": TokenType.PERCENT,
 623        "PIVOT": TokenType.PIVOT,
 624        "PRAGMA": TokenType.PRAGMA,
 625        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 626        "PROCEDURE": TokenType.PROCEDURE,
 627        "QUALIFY": TokenType.QUALIFY,
 628        "RANGE": TokenType.RANGE,
 629        "RECURSIVE": TokenType.RECURSIVE,
 630        "REGEXP": TokenType.RLIKE,
 631        "REPLACE": TokenType.REPLACE,
 632        "RETURNING": TokenType.RETURNING,
 633        "REFERENCES": TokenType.REFERENCES,
 634        "RIGHT": TokenType.RIGHT,
 635        "RLIKE": TokenType.RLIKE,
 636        "ROLLBACK": TokenType.ROLLBACK,
 637        "ROLLUP": TokenType.ROLLUP,
 638        "ROW": TokenType.ROW,
 639        "ROWS": TokenType.ROWS,
 640        "SCHEMA": TokenType.SCHEMA,
 641        "SELECT": TokenType.SELECT,
 642        "SEMI": TokenType.SEMI,
 643        "SET": TokenType.SET,
 644        "SETTINGS": TokenType.SETTINGS,
 645        "SHOW": TokenType.SHOW,
 646        "SIMILAR TO": TokenType.SIMILAR_TO,
 647        "SOME": TokenType.SOME,
 648        "SORT BY": TokenType.SORT_BY,
 649        "START WITH": TokenType.START_WITH,
 650        "TABLE": TokenType.TABLE,
 651        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 652        "TEMP": TokenType.TEMPORARY,
 653        "TEMPORARY": TokenType.TEMPORARY,
 654        "THEN": TokenType.THEN,
 655        "TRUE": TokenType.TRUE,
 656        "UNION": TokenType.UNION,
 657        "UNKNOWN": TokenType.UNKNOWN,
 658        "UNNEST": TokenType.UNNEST,
 659        "UNPIVOT": TokenType.UNPIVOT,
 660        "UPDATE": TokenType.UPDATE,
 661        "USE": TokenType.USE,
 662        "USING": TokenType.USING,
 663        "UUID": TokenType.UUID,
 664        "VALUES": TokenType.VALUES,
 665        "VIEW": TokenType.VIEW,
 666        "VOLATILE": TokenType.VOLATILE,
 667        "WHEN": TokenType.WHEN,
 668        "WHERE": TokenType.WHERE,
 669        "WINDOW": TokenType.WINDOW,
 670        "WITH": TokenType.WITH,
 671        "APPLY": TokenType.APPLY,
 672        "ARRAY": TokenType.ARRAY,
 673        "BIT": TokenType.BIT,
 674        "BOOL": TokenType.BOOLEAN,
 675        "BOOLEAN": TokenType.BOOLEAN,
 676        "BYTE": TokenType.TINYINT,
 677        "TINYINT": TokenType.TINYINT,
 678        "SHORT": TokenType.SMALLINT,
 679        "SMALLINT": TokenType.SMALLINT,
 680        "INT128": TokenType.INT128,
 681        "INT2": TokenType.SMALLINT,
 682        "INTEGER": TokenType.INT,
 683        "INT": TokenType.INT,
 684        "INT4": TokenType.INT,
 685        "LONG": TokenType.BIGINT,
 686        "BIGINT": TokenType.BIGINT,
 687        "INT8": TokenType.BIGINT,
 688        "DEC": TokenType.DECIMAL,
 689        "DECIMAL": TokenType.DECIMAL,
 690        "BIGDECIMAL": TokenType.BIGDECIMAL,
 691        "BIGNUMERIC": TokenType.BIGDECIMAL,
 692        "MAP": TokenType.MAP,
 693        "NULLABLE": TokenType.NULLABLE,
 694        "NUMBER": TokenType.DECIMAL,
 695        "NUMERIC": TokenType.DECIMAL,
 696        "FIXED": TokenType.DECIMAL,
 697        "REAL": TokenType.FLOAT,
 698        "FLOAT": TokenType.FLOAT,
 699        "FLOAT4": TokenType.FLOAT,
 700        "FLOAT8": TokenType.DOUBLE,
 701        "DOUBLE": TokenType.DOUBLE,
 702        "DOUBLE PRECISION": TokenType.DOUBLE,
 703        "JSON": TokenType.JSON,
 704        "CHAR": TokenType.CHAR,
 705        "CHARACTER": TokenType.CHAR,
 706        "NCHAR": TokenType.NCHAR,
 707        "VARCHAR": TokenType.VARCHAR,
 708        "VARCHAR2": TokenType.VARCHAR,
 709        "NVARCHAR": TokenType.NVARCHAR,
 710        "NVARCHAR2": TokenType.NVARCHAR,
 711        "STR": TokenType.TEXT,
 712        "STRING": TokenType.TEXT,
 713        "TEXT": TokenType.TEXT,
 714        "CLOB": TokenType.TEXT,
 715        "LONGVARCHAR": TokenType.TEXT,
 716        "BINARY": TokenType.BINARY,
 717        "BLOB": TokenType.VARBINARY,
 718        "BYTEA": TokenType.VARBINARY,
 719        "VARBINARY": TokenType.VARBINARY,
 720        "TIME": TokenType.TIME,
 721        "TIMETZ": TokenType.TIMETZ,
 722        "TIMESTAMP": TokenType.TIMESTAMP,
 723        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 724        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 725        "DATE": TokenType.DATE,
 726        "DATETIME": TokenType.DATETIME,
 727        "INT4RANGE": TokenType.INT4RANGE,
 728        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 729        "INT8RANGE": TokenType.INT8RANGE,
 730        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 731        "NUMRANGE": TokenType.NUMRANGE,
 732        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 733        "TSRANGE": TokenType.TSRANGE,
 734        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 735        "TSTZRANGE": TokenType.TSTZRANGE,
 736        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 737        "DATERANGE": TokenType.DATERANGE,
 738        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 739        "UNIQUE": TokenType.UNIQUE,
 740        "STRUCT": TokenType.STRUCT,
 741        "VARIANT": TokenType.VARIANT,
 742        "ALTER": TokenType.ALTER,
 743        "ANALYZE": TokenType.COMMAND,
 744        "CALL": TokenType.COMMAND,
 745        "COMMENT": TokenType.COMMENT,
 746        "COPY": TokenType.COMMAND,
 747        "EXPLAIN": TokenType.COMMAND,
 748        "GRANT": TokenType.COMMAND,
 749        "OPTIMIZE": TokenType.COMMAND,
 750        "PREPARE": TokenType.COMMAND,
 751        "TRUNCATE": TokenType.COMMAND,
 752        "VACUUM": TokenType.COMMAND,
 753        "USER-DEFINED": TokenType.USERDEFINED,
 754        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 755        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 756    }
 757
 758    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 759        " ": TokenType.SPACE,
 760        "\t": TokenType.SPACE,
 761        "\n": TokenType.BREAK,
 762        "\r": TokenType.BREAK,
 763        "\r\n": TokenType.BREAK,
 764    }
 765
 766    COMMANDS = {
 767        TokenType.COMMAND,
 768        TokenType.EXECUTE,
 769        TokenType.FETCH,
 770        TokenType.SHOW,
 771    }
 772
 773    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 774
 775    # handle numeric literals like in hive (3L = BIGINT)
 776    NUMERIC_LITERALS: t.Dict[str, str] = {}
 777    ENCODE: t.Optional[str] = None
 778
 779    COMMENTS = ["--", ("/*", "*/")]
 780
 781    __slots__ = (
 782        "sql",
 783        "size",
 784        "tokens",
 785        "_start",
 786        "_current",
 787        "_line",
 788        "_col",
 789        "_comments",
 790        "_char",
 791        "_end",
 792        "_peek",
 793        "_prev_token_line",
 794    )
 795
 796    def __init__(self) -> None:
 797        self.reset()
 798
 799    def reset(self) -> None:
 800        self.sql = ""
 801        self.size = 0
 802        self.tokens: t.List[Token] = []
 803        self._start = 0
 804        self._current = 0
 805        self._line = 1
 806        self._col = 0
 807        self._comments: t.List[str] = []
 808
 809        self._char = ""
 810        self._end = False
 811        self._peek = ""
 812        self._prev_token_line = -1
 813
 814    def tokenize(self, sql: str) -> t.List[Token]:
 815        """Returns a list of tokens corresponding to the SQL string `sql`."""
 816        self.reset()
 817        self.sql = sql
 818        self.size = len(sql)
 819
 820        try:
 821            self._scan()
 822        except Exception as e:
 823            start = max(self._current - 50, 0)
 824            end = min(self._current + 50, self.size - 1)
 825            context = self.sql[start:end]
 826            raise TokenError(f"Error tokenizing '{context}'") from e
 827
 828        return self.tokens
 829
 830    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
 831        while self.size and not self._end:
 832            self._start = self._current
 833            self._advance()
 834
 835            if self._char is None:
 836                break
 837
 838            if self._char not in self.WHITE_SPACE:
 839                if self._char.isdigit():
 840                    self._scan_number()
 841                elif self._char in self._IDENTIFIERS:
 842                    self._scan_identifier(self._IDENTIFIERS[self._char])
 843                else:
 844                    self._scan_keywords()
 845
 846            if until and until():
 847                break
 848
 849        if self.tokens and self._comments:
 850            self.tokens[-1].comments.extend(self._comments)
 851
 852    def _chars(self, size: int) -> str:
 853        if size == 1:
 854            return self._char
 855
 856        start = self._current - 1
 857        end = start + size
 858
 859        return self.sql[start:end] if end <= self.size else ""
 860
 861    def _advance(self, i: int = 1, alnum: bool = False) -> None:
 862        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
 863            self._col = 1
 864            self._line += 1
 865        else:
 866            self._col += i
 867
 868        self._current += i
 869        self._end = self._current >= self.size
 870        self._char = self.sql[self._current - 1]
 871        self._peek = "" if self._end else self.sql[self._current]
 872
 873        if alnum and self._char.isalnum():
 874            # Here we use local variables instead of attributes for better performance
 875            _col = self._col
 876            _current = self._current
 877            _end = self._end
 878            _peek = self._peek
 879
 880            while _peek.isalnum():
 881                _col += 1
 882                _current += 1
 883                _end = _current >= self.size
 884                _peek = "" if _end else self.sql[_current]
 885
 886            self._col = _col
 887            self._current = _current
 888            self._end = _end
 889            self._peek = _peek
 890            self._char = self.sql[_current - 1]
 891
 892    @property
 893    def _text(self) -> str:
 894        return self.sql[self._start : self._current]
 895
 896    def peek(self, i: int = 0) -> str:
 897        i = self._current + i
 898        if i < self.size:
 899            return self.sql[i]
 900        return ""
 901
 902    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
 903        self._prev_token_line = self._line
 904
 905        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
 906            self.tokens[-1].comments.extend(self._comments)
 907            self._comments = []
 908
 909        self.tokens.append(
 910            Token(
 911                token_type,
 912                text=self._text if text is None else text,
 913                line=self._line,
 914                col=self._col,
 915                start=self._start,
 916                end=self._current - 1,
 917                comments=self._comments,
 918            )
 919        )
 920        self._comments = []
 921
 922        # If we have either a semicolon or a begin token before the command's token, we'll parse
 923        # whatever follows the command's token as a string
 924        if (
 925            token_type in self.COMMANDS
 926            and self._peek != ";"
 927            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
 928        ):
 929            start = self._current
 930            tokens = len(self.tokens)
 931            self._scan(lambda: self._peek == ";")
 932            self.tokens = self.tokens[:tokens]
 933            text = self.sql[start : self._current].strip()
 934            if text:
 935                self._add(TokenType.STRING, text)
 936
 937    def _scan_keywords(self) -> None:
 938        size = 0
 939        word = None
 940        chars = self._text
 941        char = chars
 942        prev_space = False
 943        skip = False
 944        trie = self._KEYWORD_TRIE
 945        single_token = char in self.SINGLE_TOKENS
 946
 947        while chars:
 948            if skip:
 949                result = TrieResult.PREFIX
 950            else:
 951                result, trie = in_trie(trie, char.upper())
 952
 953            if result == TrieResult.FAILED:
 954                break
 955            if result == TrieResult.EXISTS:
 956                word = chars
 957
 958            end = self._current + size
 959            size += 1
 960
 961            if end < self.size:
 962                char = self.sql[end]
 963                single_token = single_token or char in self.SINGLE_TOKENS
 964                is_space = char in self.WHITE_SPACE
 965
 966                if not is_space or not prev_space:
 967                    if is_space:
 968                        char = " "
 969                    chars += char
 970                    prev_space = is_space
 971                    skip = False
 972                else:
 973                    skip = True
 974            else:
 975                char = ""
 976                chars = " "
 977
 978        if word:
 979            if self._scan_string(word):
 980                return
 981            if self._scan_comment(word):
 982                return
 983            if prev_space or single_token or not char:
 984                self._advance(size - 1)
 985                word = word.upper()
 986                self._add(self.KEYWORDS[word], text=word)
 987                return
 988        if self._char in self.SINGLE_TOKENS:
 989            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
 990            return
 991        self._scan_var()
 992
 993    def _scan_comment(self, comment_start: str) -> bool:
 994        if comment_start not in self._COMMENTS:
 995            return False
 996
 997        comment_start_line = self._line
 998        comment_start_size = len(comment_start)
 999        comment_end = self._COMMENTS[comment_start]
1000
1001        if comment_end:
1002            # Skip the comment's start delimiter
1003            self._advance(comment_start_size)
1004
1005            comment_end_size = len(comment_end)
1006            while not self._end and self._chars(comment_end_size) != comment_end:
1007                self._advance(alnum=True)
1008
1009            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1010            self._advance(comment_end_size - 1)
1011        else:
1012            while not self._end and not self.WHITE_SPACE.get(self._peek) is TokenType.BREAK:
1013                self._advance(alnum=True)
1014            self._comments.append(self._text[comment_start_size:])
1015
1016        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1017        # Multiple consecutive comments are preserved by appending them to the current comments list.
1018        if comment_start_line == self._prev_token_line:
1019            self.tokens[-1].comments.extend(self._comments)
1020            self._comments = []
1021            self._prev_token_line = self._line
1022
1023        return True
1024
1025    def _scan_number(self) -> None:
1026        if self._char == "0":
1027            peek = self._peek.upper()
1028            if peek == "B":
1029                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1030            elif peek == "X":
1031                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1032
1033        decimal = False
1034        scientific = 0
1035
1036        while True:
1037            if self._peek.isdigit():
1038                self._advance()
1039            elif self._peek == "." and not decimal:
1040                after = self.peek(1)
1041                if after.isdigit() or not after.isalpha():
1042                    decimal = True
1043                    self._advance()
1044                else:
1045                    return self._add(TokenType.VAR)
1046            elif self._peek in ("-", "+") and scientific == 1:
1047                scientific += 1
1048                self._advance()
1049            elif self._peek.upper() == "E" and not scientific:
1050                scientific += 1
1051                self._advance()
1052            elif self._peek.isidentifier():
1053                number_text = self._text
1054                literal = ""
1055
1056                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1057                    literal += self._peek.upper()
1058                    self._advance()
1059
1060                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal, ""))
1061
1062                if token_type:
1063                    self._add(TokenType.NUMBER, number_text)
1064                    self._add(TokenType.DCOLON, "::")
1065                    return self._add(token_type, literal)
1066                elif self.IDENTIFIERS_CAN_START_WITH_DIGIT:
1067                    return self._add(TokenType.VAR)
1068
1069                self._advance(-len(literal))
1070                return self._add(TokenType.NUMBER, number_text)
1071            else:
1072                return self._add(TokenType.NUMBER)
1073
1074    def _scan_bits(self) -> None:
1075        self._advance()
1076        value = self._extract_value()
1077        try:
1078            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1079            int(value, 2)
1080            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1081        except ValueError:
1082            self._add(TokenType.IDENTIFIER)
1083
1084    def _scan_hex(self) -> None:
1085        self._advance()
1086        value = self._extract_value()
1087        try:
1088            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1089            int(value, 16)
1090            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1091        except ValueError:
1092            self._add(TokenType.IDENTIFIER)
1093
1094    def _extract_value(self) -> str:
1095        while True:
1096            char = self._peek.strip()
1097            if char and char not in self.SINGLE_TOKENS:
1098                self._advance(alnum=True)
1099            else:
1100                break
1101
1102        return self._text
1103
1104    def _scan_string(self, start: str) -> bool:
1105        base = None
1106        token_type = TokenType.STRING
1107
1108        if start in self._QUOTES:
1109            end = self._QUOTES[start]
1110        elif start in self._FORMAT_STRINGS:
1111            end, token_type = self._FORMAT_STRINGS[start]
1112
1113            if token_type == TokenType.HEX_STRING:
1114                base = 16
1115            elif token_type == TokenType.BIT_STRING:
1116                base = 2
1117        else:
1118            return False
1119
1120        self._advance(len(start))
1121        text = self._extract_string(end)
1122
1123        if base:
1124            try:
1125                int(text, base)
1126            except:
1127                raise TokenError(
1128                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1129                )
1130        else:
1131            text = text.encode(self.ENCODE).decode(self.ENCODE) if self.ENCODE else text
1132
1133        self._add(token_type, text)
1134        return True
1135
1136    def _scan_identifier(self, identifier_end: str) -> None:
1137        self._advance()
1138        text = self._extract_string(identifier_end, self._IDENTIFIER_ESCAPES)
1139        self._add(TokenType.IDENTIFIER, text)
1140
1141    def _scan_var(self) -> None:
1142        while True:
1143            char = self._peek.strip()
1144            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1145                self._advance(alnum=True)
1146            else:
1147                break
1148
1149        self._add(
1150            TokenType.VAR
1151            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1152            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1153        )
1154
1155    def _extract_string(self, delimiter: str, escapes=None) -> str:
1156        text = ""
1157        delim_size = len(delimiter)
1158        escapes = self._STRING_ESCAPES if escapes is None else escapes
1159
1160        while True:
1161            if self._char in escapes and (self._peek == delimiter or self._peek in escapes):
1162                if self._peek == delimiter:
1163                    text += self._peek
1164                else:
1165                    text += self._char + self._peek
1166
1167                if self._current + 1 < self.size:
1168                    self._advance(2)
1169                else:
1170                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1171            else:
1172                if self._chars(delim_size) == delimiter:
1173                    if delim_size > 1:
1174                        self._advance(delim_size - 1)
1175                    break
1176
1177                if self._end:
1178                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1179
1180                current = self._current - 1
1181                self._advance(alnum=True)
1182                text += self.sql[current : self._current - 1]
1183
1184        return text