lexers/sql.lua

   1 -- Copyright 2006-2016 Mitchell mitchell.att.foicica.com. See LICENSE.
   2 -- SQL LPeg lexer.
   3
   4 local l = require('lexer')
   5 local token, word_match = l.token, l.word_match
   6 local P, R, S = lpeg.P, lpeg.R, lpeg.S
   7
   8 local M = {_NAME = 'sql'}
   9
  10 -- Whitespace.
  11 local ws = token(l.WHITESPACE, l.space^1)
  12
  13 -- Comments.
  14 local line_comment = (P('--') + '#') * l.nonnewline^0
  15 local block_comment = '/*' * (l.any - '*/')^0 * P('*/')^-1
  16 local comment = token(l.COMMENT, line_comment + block_comment)
  17
  18 -- Strings.
  19 local sq_str = l.delimited_range("'")
  20 local dq_str = l.delimited_range('"')
  21 local bt_str = l.delimited_range('`')
  22 local string = token(l.STRING, sq_str + dq_str + bt_str)
  23
  24 -- Numbers.
  25 local number = token(l.NUMBER, l.float + l.integer)
  26
  27 -- Keywords.
  28 local keyword = token(l.KEYWORD, word_match({
  29   'add', 'all', 'alter', 'analyze', 'and', 'as', 'asc', 'asensitive', 'before',
  30   'between', 'bigint', 'binary', 'blob', 'both', 'by', 'call', 'cascade',
  31   'case', 'change', 'char', 'character', 'check', 'collate', 'column',
  32   'condition', 'connection', 'constraint', 'continue', 'convert', 'create',
  33   'cross', 'current_date', 'current_time', 'current_timestamp', 'current_user',
  34   'cursor', 'database', 'databases', 'day_hour', 'day_microsecond',
  35   'day_minute', 'day_second', 'dec', 'decimal', 'declare', 'default', 'delayed',
  36   'delete', 'desc', 'describe', 'deterministic', 'distinct', 'distinctrow',
  37   'div', 'double', 'drop', 'dual', 'each', 'else', 'elseif', 'enclosed',
  38   'escaped', 'exists', 'exit', 'explain', 'false', 'fetch', 'float', 'for',
  39   'force', 'foreign', 'from', 'fulltext', 'goto', 'grant', 'group', 'having',
  40   'high_priority', 'hour_microsecond', 'hour_minute', 'hour_second', 'if',
  41   'ignore', 'in', 'index', 'infile', 'inner', 'inout', 'insensitive', 'insert',
  42   'int', 'integer', 'interval', 'into', 'is', 'iterate', 'join', 'key', 'keys',
  43   'kill', 'leading', 'leave', 'left', 'like', 'limit', 'lines', 'load',
  44   'localtime', 'localtimestamp', 'lock', 'long', 'longblob', 'longtext', 'loop',
  45   'low_priority', 'match', 'mediumblob', 'mediumint', 'mediumtext', 'middleint',
  46   'minute_microsecond', 'minute_second', 'mod', 'modifies', 'natural', 'not',
  47   'no_write_to_binlog', 'null', 'numeric', 'on', 'optimize', 'option',
  48   'optionally', 'or', 'order', 'out', 'outer', 'outfile', 'precision',
  49   'primary', 'procedure', 'purge', 'read', 'reads', 'real', 'references',
  50   'regexp', 'rename', 'repeat', 'replace', 'require', 'restrict', 'return',
  51   'revoke', 'right', 'rlike', 'schema', 'schemas', 'second_microsecond',
  52   'select', 'sensitive', 'separator', 'set', 'show', 'smallint', 'soname',
  53   'spatial', 'specific', 'sql', 'sqlexception', 'sqlstate', 'sqlwarning',
  54   'sql_big_result', 'sql_calc_found_rows', 'sql_small_result', 'ssl',
  55   'starting', 'straight_join', 'table', 'terminated', 'text', 'then',
  56   'tinyblob', 'tinyint', 'tinytext', 'to', 'trailing', 'trigger', 'true',
  57   'undo', 'union', 'unique', 'unlock', 'unsigned', 'update', 'usage', 'use',
  58   'using', 'utc_date', 'utc_time', 'utc_timestamp', 'values', 'varbinary',
  59   'varchar', 'varcharacter', 'varying', 'when', 'where', 'while', 'with',
  60   'write', 'xor', 'year_month', 'zerofill'
  61 }, nil, true))
  62
  63 -- Identifiers.
  64 local identifier = token(l.IDENTIFIER, l.word)
  65
  66 -- Operators.
  67 local operator = token(l.OPERATOR, S(',()'))
  68
  69 M._rules = {
  70   {'whitespace', ws},
  71   {'keyword', keyword},
  72   {'identifier', identifier},
  73   {'string', string},
  74   {'comment', comment},
  75   {'number', number},
  76   {'operator', operator},
  77 }
  78
  79 return M