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    HEREDOC_STRING = auto()
  81
  82    # types
  83    BIT = auto()
  84    BOOLEAN = auto()
  85    TINYINT = auto()
  86    UTINYINT = auto()
  87    SMALLINT = auto()
  88    USMALLINT = auto()
  89    MEDIUMINT = auto()
  90    UMEDIUMINT = auto()
  91    INT = auto()
  92    UINT = auto()
  93    BIGINT = auto()
  94    UBIGINT = auto()
  95    INT128 = auto()
  96    UINT128 = auto()
  97    INT256 = auto()
  98    UINT256 = auto()
  99    FLOAT = auto()
 100    DOUBLE = auto()
 101    DECIMAL = auto()
 102    UDECIMAL = auto()
 103    BIGDECIMAL = auto()
 104    CHAR = auto()
 105    NCHAR = auto()
 106    VARCHAR = auto()
 107    NVARCHAR = auto()
 108    TEXT = auto()
 109    MEDIUMTEXT = auto()
 110    LONGTEXT = auto()
 111    MEDIUMBLOB = auto()
 112    LONGBLOB = auto()
 113    TINYBLOB = auto()
 114    TINYTEXT = auto()
 115    BINARY = auto()
 116    VARBINARY = auto()
 117    JSON = auto()
 118    JSONB = auto()
 119    TIME = auto()
 120    TIMETZ = auto()
 121    TIMESTAMP = auto()
 122    TIMESTAMPTZ = auto()
 123    TIMESTAMPLTZ = auto()
 124    DATETIME = auto()
 125    DATETIME64 = auto()
 126    DATE = auto()
 127    INT4RANGE = auto()
 128    INT4MULTIRANGE = auto()
 129    INT8RANGE = auto()
 130    INT8MULTIRANGE = auto()
 131    NUMRANGE = auto()
 132    NUMMULTIRANGE = auto()
 133    TSRANGE = auto()
 134    TSMULTIRANGE = auto()
 135    TSTZRANGE = auto()
 136    TSTZMULTIRANGE = auto()
 137    DATERANGE = auto()
 138    DATEMULTIRANGE = auto()
 139    UUID = auto()
 140    GEOGRAPHY = auto()
 141    NULLABLE = auto()
 142    GEOMETRY = auto()
 143    HLLSKETCH = auto()
 144    HSTORE = auto()
 145    SUPER = auto()
 146    SERIAL = auto()
 147    SMALLSERIAL = auto()
 148    BIGSERIAL = auto()
 149    XML = auto()
 150    YEAR = auto()
 151    UNIQUEIDENTIFIER = auto()
 152    USERDEFINED = auto()
 153    MONEY = auto()
 154    SMALLMONEY = auto()
 155    ROWVERSION = auto()
 156    IMAGE = auto()
 157    VARIANT = auto()
 158    OBJECT = auto()
 159    INET = auto()
 160    IPADDRESS = auto()
 161    IPPREFIX = auto()
 162    ENUM = auto()
 163    ENUM8 = auto()
 164    ENUM16 = auto()
 165    FIXEDSTRING = auto()
 166    LOWCARDINALITY = auto()
 167    NESTED = auto()
 168    UNKNOWN = auto()
 169
 170    # keywords
 171    ALIAS = auto()
 172    ALTER = auto()
 173    ALWAYS = auto()
 174    ALL = auto()
 175    ANTI = auto()
 176    ANY = auto()
 177    APPLY = auto()
 178    ARRAY = auto()
 179    ASC = auto()
 180    ASOF = auto()
 181    AUTO_INCREMENT = auto()
 182    BEGIN = auto()
 183    BETWEEN = auto()
 184    CACHE = auto()
 185    CASE = auto()
 186    CHARACTER_SET = auto()
 187    CLUSTER_BY = auto()
 188    COLLATE = auto()
 189    COMMAND = auto()
 190    COMMENT = auto()
 191    COMMIT = auto()
 192    CONNECT_BY = auto()
 193    CONSTRAINT = auto()
 194    CREATE = auto()
 195    CROSS = auto()
 196    CUBE = auto()
 197    CURRENT_DATE = auto()
 198    CURRENT_DATETIME = auto()
 199    CURRENT_TIME = auto()
 200    CURRENT_TIMESTAMP = auto()
 201    CURRENT_USER = auto()
 202    DEFAULT = auto()
 203    DELETE = auto()
 204    DESC = auto()
 205    DESCRIBE = auto()
 206    DICTIONARY = auto()
 207    DISTINCT = auto()
 208    DISTRIBUTE_BY = auto()
 209    DIV = auto()
 210    DROP = auto()
 211    ELSE = auto()
 212    END = auto()
 213    ESCAPE = auto()
 214    EXCEPT = auto()
 215    EXECUTE = auto()
 216    EXISTS = auto()
 217    FALSE = auto()
 218    FETCH = auto()
 219    FILTER = auto()
 220    FINAL = auto()
 221    FIRST = auto()
 222    FOR = auto()
 223    FORCE = auto()
 224    FOREIGN_KEY = auto()
 225    FORMAT = auto()
 226    FROM = auto()
 227    FULL = auto()
 228    FUNCTION = auto()
 229    GLOB = auto()
 230    GLOBAL = auto()
 231    GROUP_BY = auto()
 232    GROUPING_SETS = auto()
 233    HAVING = auto()
 234    HINT = auto()
 235    IGNORE = auto()
 236    ILIKE = auto()
 237    ILIKE_ANY = auto()
 238    IN = auto()
 239    INDEX = auto()
 240    INNER = auto()
 241    INSERT = auto()
 242    INTERSECT = auto()
 243    INTERVAL = auto()
 244    INTO = auto()
 245    INTRODUCER = auto()
 246    IRLIKE = auto()
 247    IS = auto()
 248    ISNULL = auto()
 249    JOIN = auto()
 250    JOIN_MARKER = auto()
 251    KEEP = auto()
 252    KILL = auto()
 253    LANGUAGE = auto()
 254    LATERAL = auto()
 255    LEFT = auto()
 256    LIKE = auto()
 257    LIKE_ANY = auto()
 258    LIMIT = auto()
 259    LOAD = auto()
 260    LOCK = auto()
 261    MAP = auto()
 262    MATCH_RECOGNIZE = auto()
 263    MEMBER_OF = auto()
 264    MERGE = auto()
 265    MOD = auto()
 266    MODEL = auto()
 267    NATURAL = auto()
 268    NEXT = auto()
 269    NOTNULL = auto()
 270    NULL = auto()
 271    OBJECT_IDENTIFIER = auto()
 272    OFFSET = auto()
 273    ON = auto()
 274    ORDER_BY = auto()
 275    ORDERED = auto()
 276    ORDINALITY = auto()
 277    OUTER = auto()
 278    OVER = auto()
 279    OVERLAPS = auto()
 280    OVERWRITE = auto()
 281    PARTITION = auto()
 282    PARTITION_BY = auto()
 283    PERCENT = auto()
 284    PIVOT = auto()
 285    PLACEHOLDER = auto()
 286    PRAGMA = auto()
 287    PRIMARY_KEY = auto()
 288    PROCEDURE = auto()
 289    PROPERTIES = auto()
 290    PSEUDO_TYPE = auto()
 291    QUALIFY = auto()
 292    QUOTE = auto()
 293    RANGE = auto()
 294    RECURSIVE = auto()
 295    REPLACE = auto()
 296    RETURNING = auto()
 297    REFERENCES = auto()
 298    RIGHT = auto()
 299    RLIKE = auto()
 300    ROLLBACK = auto()
 301    ROLLUP = auto()
 302    ROW = auto()
 303    ROWS = auto()
 304    SELECT = auto()
 305    SEMI = auto()
 306    SEPARATOR = auto()
 307    SERDE_PROPERTIES = auto()
 308    SET = auto()
 309    SETTINGS = auto()
 310    SHOW = auto()
 311    SIMILAR_TO = auto()
 312    SOME = auto()
 313    SORT_BY = auto()
 314    START_WITH = auto()
 315    STRUCT = auto()
 316    TABLE_SAMPLE = auto()
 317    TEMPORARY = auto()
 318    TOP = auto()
 319    THEN = auto()
 320    TRUE = auto()
 321    UNCACHE = auto()
 322    UNION = auto()
 323    UNNEST = auto()
 324    UNPIVOT = auto()
 325    UPDATE = auto()
 326    USE = auto()
 327    USING = auto()
 328    VALUES = auto()
 329    VIEW = auto()
 330    VOLATILE = auto()
 331    WHEN = auto()
 332    WHERE = auto()
 333    WINDOW = auto()
 334    WITH = auto()
 335    UNIQUE = auto()
 336    VERSION_SNAPSHOT = auto()
 337    TIMESTAMP_SNAPSHOT = auto()
 338
 339
 340class Token:
 341    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 342
 343    @classmethod
 344    def number(cls, number: int) -> Token:
 345        """Returns a NUMBER token with `number` as its text."""
 346        return cls(TokenType.NUMBER, str(number))
 347
 348    @classmethod
 349    def string(cls, string: str) -> Token:
 350        """Returns a STRING token with `string` as its text."""
 351        return cls(TokenType.STRING, string)
 352
 353    @classmethod
 354    def identifier(cls, identifier: str) -> Token:
 355        """Returns an IDENTIFIER token with `identifier` as its text."""
 356        return cls(TokenType.IDENTIFIER, identifier)
 357
 358    @classmethod
 359    def var(cls, var: str) -> Token:
 360        """Returns an VAR token with `var` as its text."""
 361        return cls(TokenType.VAR, var)
 362
 363    def __init__(
 364        self,
 365        token_type: TokenType,
 366        text: str,
 367        line: int = 1,
 368        col: int = 1,
 369        start: int = 0,
 370        end: int = 0,
 371        comments: t.List[str] = [],
 372    ) -> None:
 373        """Token initializer.
 374
 375        Args:
 376            token_type: The TokenType Enum.
 377            text: The text of the token.
 378            line: The line that the token ends on.
 379            col: The column that the token ends on.
 380            start: The start index of the token.
 381            end: The ending index of the token.
 382            comments: The comments to attach to the token.
 383        """
 384        self.token_type = token_type
 385        self.text = text
 386        self.line = line
 387        self.col = col
 388        self.start = start
 389        self.end = end
 390        self.comments = comments
 391
 392    def __repr__(self) -> str:
 393        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 394        return f"<Token {attributes}>"
 395
 396
 397class _Tokenizer(type):
 398    def __new__(cls, clsname, bases, attrs):
 399        klass = super().__new__(cls, clsname, bases, attrs)
 400
 401        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 402            return dict(
 403                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 404            )
 405
 406        def _quotes_to_format(
 407            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 408        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 409            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 410
 411        klass._QUOTES = _convert_quotes(klass.QUOTES)
 412        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 413
 414        klass._FORMAT_STRINGS = {
 415            **{
 416                p + s: (e, TokenType.NATIONAL_STRING)
 417                for s, e in klass._QUOTES.items()
 418                for p in ("n", "N")
 419            },
 420            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 421            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 422            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 423            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 424            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 425        }
 426
 427        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 428        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 429        klass._COMMENTS = {
 430            **dict(
 431                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 432                for comment in klass.COMMENTS
 433            ),
 434            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 435        }
 436
 437        klass._KEYWORD_TRIE = new_trie(
 438            key.upper()
 439            for key in (
 440                *klass.KEYWORDS,
 441                *klass._COMMENTS,
 442                *klass._QUOTES,
 443                *klass._FORMAT_STRINGS,
 444            )
 445            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 446        )
 447
 448        return klass
 449
 450
 451class Tokenizer(metaclass=_Tokenizer):
 452    SINGLE_TOKENS = {
 453        "(": TokenType.L_PAREN,
 454        ")": TokenType.R_PAREN,
 455        "[": TokenType.L_BRACKET,
 456        "]": TokenType.R_BRACKET,
 457        "{": TokenType.L_BRACE,
 458        "}": TokenType.R_BRACE,
 459        "&": TokenType.AMP,
 460        "^": TokenType.CARET,
 461        ":": TokenType.COLON,
 462        ",": TokenType.COMMA,
 463        ".": TokenType.DOT,
 464        "-": TokenType.DASH,
 465        "=": TokenType.EQ,
 466        ">": TokenType.GT,
 467        "<": TokenType.LT,
 468        "%": TokenType.MOD,
 469        "!": TokenType.NOT,
 470        "|": TokenType.PIPE,
 471        "+": TokenType.PLUS,
 472        ";": TokenType.SEMICOLON,
 473        "/": TokenType.SLASH,
 474        "\\": TokenType.BACKSLASH,
 475        "*": TokenType.STAR,
 476        "~": TokenType.TILDA,
 477        "?": TokenType.PLACEHOLDER,
 478        "@": TokenType.PARAMETER,
 479        # used for breaking a var like x'y' but nothing else
 480        # the token type doesn't matter
 481        "'": TokenType.QUOTE,
 482        "`": TokenType.IDENTIFIER,
 483        '"': TokenType.IDENTIFIER,
 484        "#": TokenType.HASH,
 485    }
 486
 487    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 488    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 489    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 490    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 491    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 492    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 493    IDENTIFIER_ESCAPES = ['"']
 494    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 495    STRING_ESCAPES = ["'"]
 496    VAR_SINGLE_TOKENS: t.Set[str] = set()
 497    ESCAPE_SEQUENCES: t.Dict[str, str] = {}
 498
 499    # Autofilled
 500    IDENTIFIERS_CAN_START_WITH_DIGIT: bool = False
 501
 502    _COMMENTS: t.Dict[str, str] = {}
 503    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 504    _IDENTIFIERS: t.Dict[str, str] = {}
 505    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 506    _QUOTES: t.Dict[str, str] = {}
 507    _STRING_ESCAPES: t.Set[str] = set()
 508    _KEYWORD_TRIE: t.Dict = {}
 509
 510    KEYWORDS: t.Dict[str, TokenType] = {
 511        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 512        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 513        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 514        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 515        "/*+": TokenType.HINT,
 516        "==": TokenType.EQ,
 517        "::": TokenType.DCOLON,
 518        "||": TokenType.DPIPE,
 519        ">=": TokenType.GTE,
 520        "<=": TokenType.LTE,
 521        "<>": TokenType.NEQ,
 522        "!=": TokenType.NEQ,
 523        "<=>": TokenType.NULLSAFE_EQ,
 524        "->": TokenType.ARROW,
 525        "->>": TokenType.DARROW,
 526        "=>": TokenType.FARROW,
 527        "#>": TokenType.HASH_ARROW,
 528        "#>>": TokenType.DHASH_ARROW,
 529        "<->": TokenType.LR_ARROW,
 530        "&&": TokenType.DAMP,
 531        "??": TokenType.DQMARK,
 532        "ALL": TokenType.ALL,
 533        "ALWAYS": TokenType.ALWAYS,
 534        "AND": TokenType.AND,
 535        "ANTI": TokenType.ANTI,
 536        "ANY": TokenType.ANY,
 537        "ASC": TokenType.ASC,
 538        "AS": TokenType.ALIAS,
 539        "ASOF": TokenType.ASOF,
 540        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 541        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 542        "BEGIN": TokenType.BEGIN,
 543        "BETWEEN": TokenType.BETWEEN,
 544        "CACHE": TokenType.CACHE,
 545        "UNCACHE": TokenType.UNCACHE,
 546        "CASE": TokenType.CASE,
 547        "CHARACTER SET": TokenType.CHARACTER_SET,
 548        "CLUSTER BY": TokenType.CLUSTER_BY,
 549        "COLLATE": TokenType.COLLATE,
 550        "COLUMN": TokenType.COLUMN,
 551        "COMMIT": TokenType.COMMIT,
 552        "CONNECT BY": TokenType.CONNECT_BY,
 553        "CONSTRAINT": TokenType.CONSTRAINT,
 554        "CREATE": TokenType.CREATE,
 555        "CROSS": TokenType.CROSS,
 556        "CUBE": TokenType.CUBE,
 557        "CURRENT_DATE": TokenType.CURRENT_DATE,
 558        "CURRENT_TIME": TokenType.CURRENT_TIME,
 559        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 560        "CURRENT_USER": TokenType.CURRENT_USER,
 561        "DATABASE": TokenType.DATABASE,
 562        "DEFAULT": TokenType.DEFAULT,
 563        "DELETE": TokenType.DELETE,
 564        "DESC": TokenType.DESC,
 565        "DESCRIBE": TokenType.DESCRIBE,
 566        "DISTINCT": TokenType.DISTINCT,
 567        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 568        "DIV": TokenType.DIV,
 569        "DROP": TokenType.DROP,
 570        "ELSE": TokenType.ELSE,
 571        "END": TokenType.END,
 572        "ESCAPE": TokenType.ESCAPE,
 573        "EXCEPT": TokenType.EXCEPT,
 574        "EXECUTE": TokenType.EXECUTE,
 575        "EXISTS": TokenType.EXISTS,
 576        "FALSE": TokenType.FALSE,
 577        "FETCH": TokenType.FETCH,
 578        "FILTER": TokenType.FILTER,
 579        "FIRST": TokenType.FIRST,
 580        "FULL": TokenType.FULL,
 581        "FUNCTION": TokenType.FUNCTION,
 582        "FOR": TokenType.FOR,
 583        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 584        "FORMAT": TokenType.FORMAT,
 585        "FROM": TokenType.FROM,
 586        "GEOGRAPHY": TokenType.GEOGRAPHY,
 587        "GEOMETRY": TokenType.GEOMETRY,
 588        "GLOB": TokenType.GLOB,
 589        "GROUP BY": TokenType.GROUP_BY,
 590        "GROUPING SETS": TokenType.GROUPING_SETS,
 591        "HAVING": TokenType.HAVING,
 592        "ILIKE": TokenType.ILIKE,
 593        "IN": TokenType.IN,
 594        "INDEX": TokenType.INDEX,
 595        "INET": TokenType.INET,
 596        "INNER": TokenType.INNER,
 597        "INSERT": TokenType.INSERT,
 598        "INTERVAL": TokenType.INTERVAL,
 599        "INTERSECT": TokenType.INTERSECT,
 600        "INTO": TokenType.INTO,
 601        "IS": TokenType.IS,
 602        "ISNULL": TokenType.ISNULL,
 603        "JOIN": TokenType.JOIN,
 604        "KEEP": TokenType.KEEP,
 605        "KILL": TokenType.KILL,
 606        "LATERAL": TokenType.LATERAL,
 607        "LEFT": TokenType.LEFT,
 608        "LIKE": TokenType.LIKE,
 609        "LIMIT": TokenType.LIMIT,
 610        "LOAD": TokenType.LOAD,
 611        "LOCK": TokenType.LOCK,
 612        "MERGE": TokenType.MERGE,
 613        "NATURAL": TokenType.NATURAL,
 614        "NEXT": TokenType.NEXT,
 615        "NOT": TokenType.NOT,
 616        "NOTNULL": TokenType.NOTNULL,
 617        "NULL": TokenType.NULL,
 618        "OBJECT": TokenType.OBJECT,
 619        "OFFSET": TokenType.OFFSET,
 620        "ON": TokenType.ON,
 621        "OR": TokenType.OR,
 622        "XOR": TokenType.XOR,
 623        "ORDER BY": TokenType.ORDER_BY,
 624        "ORDINALITY": TokenType.ORDINALITY,
 625        "OUTER": TokenType.OUTER,
 626        "OVER": TokenType.OVER,
 627        "OVERLAPS": TokenType.OVERLAPS,
 628        "OVERWRITE": TokenType.OVERWRITE,
 629        "PARTITION": TokenType.PARTITION,
 630        "PARTITION BY": TokenType.PARTITION_BY,
 631        "PARTITIONED BY": TokenType.PARTITION_BY,
 632        "PARTITIONED_BY": TokenType.PARTITION_BY,
 633        "PERCENT": TokenType.PERCENT,
 634        "PIVOT": TokenType.PIVOT,
 635        "PRAGMA": TokenType.PRAGMA,
 636        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 637        "PROCEDURE": TokenType.PROCEDURE,
 638        "QUALIFY": TokenType.QUALIFY,
 639        "RANGE": TokenType.RANGE,
 640        "RECURSIVE": TokenType.RECURSIVE,
 641        "REGEXP": TokenType.RLIKE,
 642        "REPLACE": TokenType.REPLACE,
 643        "RETURNING": TokenType.RETURNING,
 644        "REFERENCES": TokenType.REFERENCES,
 645        "RIGHT": TokenType.RIGHT,
 646        "RLIKE": TokenType.RLIKE,
 647        "ROLLBACK": TokenType.ROLLBACK,
 648        "ROLLUP": TokenType.ROLLUP,
 649        "ROW": TokenType.ROW,
 650        "ROWS": TokenType.ROWS,
 651        "SCHEMA": TokenType.SCHEMA,
 652        "SELECT": TokenType.SELECT,
 653        "SEMI": TokenType.SEMI,
 654        "SET": TokenType.SET,
 655        "SETTINGS": TokenType.SETTINGS,
 656        "SHOW": TokenType.SHOW,
 657        "SIMILAR TO": TokenType.SIMILAR_TO,
 658        "SOME": TokenType.SOME,
 659        "SORT BY": TokenType.SORT_BY,
 660        "START WITH": TokenType.START_WITH,
 661        "TABLE": TokenType.TABLE,
 662        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 663        "TEMP": TokenType.TEMPORARY,
 664        "TEMPORARY": TokenType.TEMPORARY,
 665        "THEN": TokenType.THEN,
 666        "TRUE": TokenType.TRUE,
 667        "UNION": TokenType.UNION,
 668        "UNKNOWN": TokenType.UNKNOWN,
 669        "UNNEST": TokenType.UNNEST,
 670        "UNPIVOT": TokenType.UNPIVOT,
 671        "UPDATE": TokenType.UPDATE,
 672        "USE": TokenType.USE,
 673        "USING": TokenType.USING,
 674        "UUID": TokenType.UUID,
 675        "VALUES": TokenType.VALUES,
 676        "VIEW": TokenType.VIEW,
 677        "VOLATILE": TokenType.VOLATILE,
 678        "WHEN": TokenType.WHEN,
 679        "WHERE": TokenType.WHERE,
 680        "WINDOW": TokenType.WINDOW,
 681        "WITH": TokenType.WITH,
 682        "APPLY": TokenType.APPLY,
 683        "ARRAY": TokenType.ARRAY,
 684        "BIT": TokenType.BIT,
 685        "BOOL": TokenType.BOOLEAN,
 686        "BOOLEAN": TokenType.BOOLEAN,
 687        "BYTE": TokenType.TINYINT,
 688        "MEDIUMINT": TokenType.MEDIUMINT,
 689        "TINYINT": TokenType.TINYINT,
 690        "SHORT": TokenType.SMALLINT,
 691        "SMALLINT": TokenType.SMALLINT,
 692        "INT128": TokenType.INT128,
 693        "INT2": TokenType.SMALLINT,
 694        "INTEGER": TokenType.INT,
 695        "INT": TokenType.INT,
 696        "INT4": TokenType.INT,
 697        "LONG": TokenType.BIGINT,
 698        "BIGINT": TokenType.BIGINT,
 699        "INT8": TokenType.BIGINT,
 700        "DEC": TokenType.DECIMAL,
 701        "DECIMAL": TokenType.DECIMAL,
 702        "BIGDECIMAL": TokenType.BIGDECIMAL,
 703        "BIGNUMERIC": TokenType.BIGDECIMAL,
 704        "MAP": TokenType.MAP,
 705        "NULLABLE": TokenType.NULLABLE,
 706        "NUMBER": TokenType.DECIMAL,
 707        "NUMERIC": TokenType.DECIMAL,
 708        "FIXED": TokenType.DECIMAL,
 709        "REAL": TokenType.FLOAT,
 710        "FLOAT": TokenType.FLOAT,
 711        "FLOAT4": TokenType.FLOAT,
 712        "FLOAT8": TokenType.DOUBLE,
 713        "DOUBLE": TokenType.DOUBLE,
 714        "DOUBLE PRECISION": TokenType.DOUBLE,
 715        "JSON": TokenType.JSON,
 716        "CHAR": TokenType.CHAR,
 717        "CHARACTER": TokenType.CHAR,
 718        "NCHAR": TokenType.NCHAR,
 719        "VARCHAR": TokenType.VARCHAR,
 720        "VARCHAR2": TokenType.VARCHAR,
 721        "NVARCHAR": TokenType.NVARCHAR,
 722        "NVARCHAR2": TokenType.NVARCHAR,
 723        "STR": TokenType.TEXT,
 724        "STRING": TokenType.TEXT,
 725        "TEXT": TokenType.TEXT,
 726        "LONGTEXT": TokenType.LONGTEXT,
 727        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 728        "TINYTEXT": TokenType.TINYTEXT,
 729        "CLOB": TokenType.TEXT,
 730        "LONGVARCHAR": TokenType.TEXT,
 731        "BINARY": TokenType.BINARY,
 732        "BLOB": TokenType.VARBINARY,
 733        "LONGBLOB": TokenType.LONGBLOB,
 734        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 735        "TINYBLOB": TokenType.TINYBLOB,
 736        "BYTEA": TokenType.VARBINARY,
 737        "VARBINARY": TokenType.VARBINARY,
 738        "TIME": TokenType.TIME,
 739        "TIMETZ": TokenType.TIMETZ,
 740        "TIMESTAMP": TokenType.TIMESTAMP,
 741        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 742        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 743        "DATE": TokenType.DATE,
 744        "DATETIME": TokenType.DATETIME,
 745        "INT4RANGE": TokenType.INT4RANGE,
 746        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 747        "INT8RANGE": TokenType.INT8RANGE,
 748        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 749        "NUMRANGE": TokenType.NUMRANGE,
 750        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 751        "TSRANGE": TokenType.TSRANGE,
 752        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 753        "TSTZRANGE": TokenType.TSTZRANGE,
 754        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 755        "DATERANGE": TokenType.DATERANGE,
 756        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 757        "UNIQUE": TokenType.UNIQUE,
 758        "STRUCT": TokenType.STRUCT,
 759        "VARIANT": TokenType.VARIANT,
 760        "ALTER": TokenType.ALTER,
 761        "ANALYZE": TokenType.COMMAND,
 762        "CALL": TokenType.COMMAND,
 763        "COMMENT": TokenType.COMMENT,
 764        "COPY": TokenType.COMMAND,
 765        "EXPLAIN": TokenType.COMMAND,
 766        "GRANT": TokenType.COMMAND,
 767        "OPTIMIZE": TokenType.COMMAND,
 768        "PREPARE": TokenType.COMMAND,
 769        "TRUNCATE": TokenType.COMMAND,
 770        "VACUUM": TokenType.COMMAND,
 771        "USER-DEFINED": TokenType.USERDEFINED,
 772        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 773        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 774    }
 775
 776    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 777        " ": TokenType.SPACE,
 778        "\t": TokenType.SPACE,
 779        "\n": TokenType.BREAK,
 780        "\r": TokenType.BREAK,
 781        "\r\n": TokenType.BREAK,
 782    }
 783
 784    COMMANDS = {
 785        TokenType.COMMAND,
 786        TokenType.EXECUTE,
 787        TokenType.FETCH,
 788        TokenType.SHOW,
 789    }
 790
 791    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 792
 793    # handle numeric literals like in hive (3L = BIGINT)
 794    NUMERIC_LITERALS: t.Dict[str, str] = {}
 795    ENCODE: t.Optional[str] = None
 796
 797    COMMENTS = ["--", ("/*", "*/")]
 798
 799    __slots__ = (
 800        "sql",
 801        "size",
 802        "tokens",
 803        "_start",
 804        "_current",
 805        "_line",
 806        "_col",
 807        "_comments",
 808        "_char",
 809        "_end",
 810        "_peek",
 811        "_prev_token_line",
 812    )
 813
 814    def __init__(self) -> None:
 815        self.reset()
 816
 817    def reset(self) -> None:
 818        self.sql = ""
 819        self.size = 0
 820        self.tokens: t.List[Token] = []
 821        self._start = 0
 822        self._current = 0
 823        self._line = 1
 824        self._col = 0
 825        self._comments: t.List[str] = []
 826
 827        self._char = ""
 828        self._end = False
 829        self._peek = ""
 830        self._prev_token_line = -1
 831
 832    def tokenize(self, sql: str) -> t.List[Token]:
 833        """Returns a list of tokens corresponding to the SQL string `sql`."""
 834        self.reset()
 835        self.sql = sql
 836        self.size = len(sql)
 837
 838        try:
 839            self._scan()
 840        except Exception as e:
 841            start = max(self._current - 50, 0)
 842            end = min(self._current + 50, self.size - 1)
 843            context = self.sql[start:end]
 844            raise TokenError(f"Error tokenizing '{context}'") from e
 845
 846        return self.tokens
 847
 848    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
 849        while self.size and not self._end:
 850            self._start = self._current
 851            self._advance()
 852
 853            if self._char is None:
 854                break
 855
 856            if self._char not in self.WHITE_SPACE:
 857                if self._char.isdigit():
 858                    self._scan_number()
 859                elif self._char in self._IDENTIFIERS:
 860                    self._scan_identifier(self._IDENTIFIERS[self._char])
 861                else:
 862                    self._scan_keywords()
 863
 864            if until and until():
 865                break
 866
 867        if self.tokens and self._comments:
 868            self.tokens[-1].comments.extend(self._comments)
 869
 870    def _chars(self, size: int) -> str:
 871        if size == 1:
 872            return self._char
 873
 874        start = self._current - 1
 875        end = start + size
 876
 877        return self.sql[start:end] if end <= self.size else ""
 878
 879    def _advance(self, i: int = 1, alnum: bool = False) -> None:
 880        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
 881            self._col = 1
 882            self._line += 1
 883        else:
 884            self._col += i
 885
 886        self._current += i
 887        self._end = self._current >= self.size
 888        self._char = self.sql[self._current - 1]
 889        self._peek = "" if self._end else self.sql[self._current]
 890
 891        if alnum and self._char.isalnum():
 892            # Here we use local variables instead of attributes for better performance
 893            _col = self._col
 894            _current = self._current
 895            _end = self._end
 896            _peek = self._peek
 897
 898            while _peek.isalnum():
 899                _col += 1
 900                _current += 1
 901                _end = _current >= self.size
 902                _peek = "" if _end else self.sql[_current]
 903
 904            self._col = _col
 905            self._current = _current
 906            self._end = _end
 907            self._peek = _peek
 908            self._char = self.sql[_current - 1]
 909
 910    @property
 911    def _text(self) -> str:
 912        return self.sql[self._start : self._current]
 913
 914    def peek(self, i: int = 0) -> str:
 915        i = self._current + i
 916        if i < self.size:
 917            return self.sql[i]
 918        return ""
 919
 920    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
 921        self._prev_token_line = self._line
 922
 923        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
 924            self.tokens[-1].comments.extend(self._comments)
 925            self._comments = []
 926
 927        self.tokens.append(
 928            Token(
 929                token_type,
 930                text=self._text if text is None else text,
 931                line=self._line,
 932                col=self._col,
 933                start=self._start,
 934                end=self._current - 1,
 935                comments=self._comments,
 936            )
 937        )
 938        self._comments = []
 939
 940        # If we have either a semicolon or a begin token before the command's token, we'll parse
 941        # whatever follows the command's token as a string
 942        if (
 943            token_type in self.COMMANDS
 944            and self._peek != ";"
 945            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
 946        ):
 947            start = self._current
 948            tokens = len(self.tokens)
 949            self._scan(lambda: self._peek == ";")
 950            self.tokens = self.tokens[:tokens]
 951            text = self.sql[start : self._current].strip()
 952            if text:
 953                self._add(TokenType.STRING, text)
 954
 955    def _scan_keywords(self) -> None:
 956        size = 0
 957        word = None
 958        chars = self._text
 959        char = chars
 960        prev_space = False
 961        skip = False
 962        trie = self._KEYWORD_TRIE
 963        single_token = char in self.SINGLE_TOKENS
 964
 965        while chars:
 966            if skip:
 967                result = TrieResult.PREFIX
 968            else:
 969                result, trie = in_trie(trie, char.upper())
 970
 971            if result == TrieResult.FAILED:
 972                break
 973            if result == TrieResult.EXISTS:
 974                word = chars
 975
 976            end = self._current + size
 977            size += 1
 978
 979            if end < self.size:
 980                char = self.sql[end]
 981                single_token = single_token or char in self.SINGLE_TOKENS
 982                is_space = char in self.WHITE_SPACE
 983
 984                if not is_space or not prev_space:
 985                    if is_space:
 986                        char = " "
 987                    chars += char
 988                    prev_space = is_space
 989                    skip = False
 990                else:
 991                    skip = True
 992            else:
 993                char = ""
 994                chars = " "
 995
 996        if word:
 997            if self._scan_string(word):
 998                return
 999            if self._scan_comment(word):
1000                return
1001            if prev_space or single_token or not char:
1002                self._advance(size - 1)
1003                word = word.upper()
1004                self._add(self.KEYWORDS[word], text=word)
1005                return
1006
1007        if self._char in self.SINGLE_TOKENS:
1008            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1009            return
1010
1011        self._scan_var()
1012
1013    def _scan_comment(self, comment_start: str) -> bool:
1014        if comment_start not in self._COMMENTS:
1015            return False
1016
1017        comment_start_line = self._line
1018        comment_start_size = len(comment_start)
1019        comment_end = self._COMMENTS[comment_start]
1020
1021        if comment_end:
1022            # Skip the comment's start delimiter
1023            self._advance(comment_start_size)
1024
1025            comment_end_size = len(comment_end)
1026            while not self._end and self._chars(comment_end_size) != comment_end:
1027                self._advance(alnum=True)
1028
1029            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1030            self._advance(comment_end_size - 1)
1031        else:
1032            while not self._end and not self.WHITE_SPACE.get(self._peek) is TokenType.BREAK:
1033                self._advance(alnum=True)
1034            self._comments.append(self._text[comment_start_size:])
1035
1036        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1037        # Multiple consecutive comments are preserved by appending them to the current comments list.
1038        if comment_start_line == self._prev_token_line:
1039            self.tokens[-1].comments.extend(self._comments)
1040            self._comments = []
1041            self._prev_token_line = self._line
1042
1043        return True
1044
1045    def _scan_number(self) -> None:
1046        if self._char == "0":
1047            peek = self._peek.upper()
1048            if peek == "B":
1049                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1050            elif peek == "X":
1051                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1052
1053        decimal = False
1054        scientific = 0
1055
1056        while True:
1057            if self._peek.isdigit():
1058                self._advance()
1059            elif self._peek == "." and not decimal:
1060                after = self.peek(1)
1061                if after.isdigit() or not after.isalpha():
1062                    decimal = True
1063                    self._advance()
1064                else:
1065                    return self._add(TokenType.VAR)
1066            elif self._peek in ("-", "+") and scientific == 1:
1067                scientific += 1
1068                self._advance()
1069            elif self._peek.upper() == "E" and not scientific:
1070                scientific += 1
1071                self._advance()
1072            elif self._peek.isidentifier():
1073                number_text = self._text
1074                literal = ""
1075
1076                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1077                    literal += self._peek.upper()
1078                    self._advance()
1079
1080                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal, ""))
1081
1082                if token_type:
1083                    self._add(TokenType.NUMBER, number_text)
1084                    self._add(TokenType.DCOLON, "::")
1085                    return self._add(token_type, literal)
1086                elif self.IDENTIFIERS_CAN_START_WITH_DIGIT:
1087                    return self._add(TokenType.VAR)
1088
1089                self._advance(-len(literal))
1090                return self._add(TokenType.NUMBER, number_text)
1091            else:
1092                return self._add(TokenType.NUMBER)
1093
1094    def _scan_bits(self) -> None:
1095        self._advance()
1096        value = self._extract_value()
1097        try:
1098            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1099            int(value, 2)
1100            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1101        except ValueError:
1102            self._add(TokenType.IDENTIFIER)
1103
1104    def _scan_hex(self) -> None:
1105        self._advance()
1106        value = self._extract_value()
1107        try:
1108            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1109            int(value, 16)
1110            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1111        except ValueError:
1112            self._add(TokenType.IDENTIFIER)
1113
1114    def _extract_value(self) -> str:
1115        while True:
1116            char = self._peek.strip()
1117            if char and char not in self.SINGLE_TOKENS:
1118                self._advance(alnum=True)
1119            else:
1120                break
1121
1122        return self._text
1123
1124    def _scan_string(self, start: str) -> bool:
1125        base = None
1126        token_type = TokenType.STRING
1127
1128        if start in self._QUOTES:
1129            end = self._QUOTES[start]
1130        elif start in self._FORMAT_STRINGS:
1131            end, token_type = self._FORMAT_STRINGS[start]
1132
1133            if token_type == TokenType.HEX_STRING:
1134                base = 16
1135            elif token_type == TokenType.BIT_STRING:
1136                base = 2
1137            elif token_type == TokenType.HEREDOC_STRING:
1138                self._advance()
1139                tag = "" if self._char == end else self._extract_string(end)
1140                end = f"{start}{tag}{end}"
1141        else:
1142            return False
1143
1144        self._advance(len(start))
1145        text = self._extract_string(end)
1146
1147        if base:
1148            try:
1149                int(text, base)
1150            except:
1151                raise TokenError(
1152                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1153                )
1154        else:
1155            text = text.encode(self.ENCODE).decode(self.ENCODE) if self.ENCODE else text
1156
1157        self._add(token_type, text)
1158        return True
1159
1160    def _scan_identifier(self, identifier_end: str) -> None:
1161        self._advance()
1162        text = self._extract_string(identifier_end, self._IDENTIFIER_ESCAPES)
1163        self._add(TokenType.IDENTIFIER, text)
1164
1165    def _scan_var(self) -> None:
1166        while True:
1167            char = self._peek.strip()
1168            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1169                self._advance(alnum=True)
1170            else:
1171                break
1172
1173        self._add(
1174            TokenType.VAR
1175            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1176            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1177        )
1178
1179    def _extract_string(self, delimiter: str, escapes=None) -> str:
1180        text = ""
1181        delim_size = len(delimiter)
1182        escapes = self._STRING_ESCAPES if escapes is None else escapes
1183
1184        while True:
1185            if (
1186                self._char in escapes
1187                and (self._peek == delimiter or self._peek in escapes)
1188                and (self._char not in self._QUOTES or self._char == self._peek)
1189            ):
1190                if self._peek == delimiter:
1191                    text += self._peek
1192                else:
1193                    text += self._char + self._peek
1194
1195                if self._current + 1 < self.size:
1196                    self._advance(2)
1197                else:
1198                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1199            else:
1200                if self._chars(delim_size) == delimiter:
1201                    if delim_size > 1:
1202                        self._advance(delim_size - 1)
1203                    break
1204
1205                if self._end:
1206                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1207
1208                if self.ESCAPE_SEQUENCES and self._peek and self._char in self.STRING_ESCAPES:
1209                    escaped_sequence = self.ESCAPE_SEQUENCES.get(self._char + self._peek)
1210                    if escaped_sequence:
1211                        self._advance(2)
1212                        text += escaped_sequence
1213                        continue
1214
1215                current = self._current - 1
1216                self._advance(alnum=True)
1217                text += self.sql[current : self._current - 1]
1218
1219        return text