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