tartrazine/lexers/pig.xml

<lexer>
  <config>
    <name>Pig</name>
    <alias>pig</alias>
    <filename>*.pig</filename>
    <mime_type>text/x-pig</mime_type>
    <case_insensitive>true</case_insensitive>
  </config>
  <rules>
    <state name="root">
      <rule pattern="\s+">
        <token type="Text"/>
      </rule>
      <rule pattern="--.*">
        <token type="Comment"/>
      </rule>
      <rule pattern="/\*[\w\W]*?\*/">
        <token type="CommentMultiline"/>
      </rule>
      <rule pattern="\\\n">
        <token type="Text"/>
      </rule>
      <rule pattern="\\">
        <token type="Text"/>
      </rule>
      <rule pattern="\&#39;(?:\\[ntbrf\\\&#39;]|\\u[0-9a-f]{4}|[^\&#39;\\\n\r])*\&#39;">
        <token type="LiteralString"/>
      </rule>
      <rule>
        <include state="keywords"/>
      </rule>
      <rule>
        <include state="types"/>
      </rule>
      <rule>
        <include state="builtins"/>
      </rule>
      <rule>
        <include state="punct"/>
      </rule>
      <rule>
        <include state="operators"/>
      </rule>
      <rule pattern="[0-9]*\.[0-9]+(e[0-9]+)?[fd]?">
        <token type="LiteralNumberFloat"/>
      </rule>
      <rule pattern="0x[0-9a-f]+">
        <token type="LiteralNumberHex"/>
      </rule>
      <rule pattern="[0-9]+L?">
        <token type="LiteralNumberInteger"/>
      </rule>
      <rule pattern="\n">
        <token type="Text"/>
      </rule>
      <rule pattern="([a-z_]\w*)(\s*)(\()">
        <bygroups>
          <token type="NameFunction"/>
          <token type="Text"/>
          <token type="Punctuation"/>
        </bygroups>
      </rule>
      <rule pattern="[()#:]">
        <token type="Text"/>
      </rule>
      <rule pattern="[^(:#\&#39;&#34;)\s]+">
        <token type="Text"/>
      </rule>
      <rule pattern="\S+\s+">
        <token type="Text"/>
      </rule>
    </state>
    <state name="keywords">
      <rule pattern="(assert|and|any|all|arrange|as|asc|bag|by|cache|CASE|cat|cd|cp|%declare|%default|define|dense|desc|describe|distinct|du|dump|eval|exex|explain|filter|flatten|foreach|full|generate|group|help|if|illustrate|import|inner|input|into|is|join|kill|left|limit|load|ls|map|matches|mkdir|mv|not|null|onschema|or|order|outer|output|parallel|pig|pwd|quit|register|returns|right|rm|rmf|rollup|run|sample|set|ship|split|stderr|stdin|stdout|store|stream|through|union|using|void)\b">
        <token type="Keyword"/>
      </rule>
    </state>
    <state name="builtins">
      <rule pattern="(AVG|BinStorage|cogroup|CONCAT|copyFromLocal|copyToLocal|COUNT|cross|DIFF|MAX|MIN|PigDump|PigStorage|SIZE|SUM|TextLoader|TOKENIZE)\b">
        <token type="NameBuiltin"/>
      </rule>
    </state>
    <state name="types">
      <rule pattern="(bytearray|BIGINTEGER|BIGDECIMAL|chararray|datetime|double|float|int|long|tuple)\b">
        <token type="KeywordType"/>
      </rule>
    </state>
    <state name="punct">
      <rule pattern="[;(){}\[\]]">
        <token type="Punctuation"/>
      </rule>
    </state>
    <state name="operators">
      <rule pattern="[#=,./%+\-?]">
        <token type="Operator"/>
      </rule>
      <rule pattern="(eq|gt|lt|gte|lte|neq|matches)\b">
        <token type="Operator"/>
      </rule>
      <rule pattern="(==|&lt;=|&lt;|&gt;=|&gt;|!=)">
        <token type="Operator"/>
      </rule>
    </state>
  </rules>
</lexer>