Full text search syntax and semantics

XMLWordPrintable

    • Type: Bug
    • Resolution: Fixed
    • Priority: Major
    • None
    • Affects Version/s: None
    • Component/s: Domain Model
    • None
    • Hide

      As defined by JSR-283, but modified to the style of the CMIS SQL grammar.

      Note:
      <space> & <non space char> definitions taken directly from JSR-283, but consideration of whitespace is probably needed.

      <text search expression> ::= <disjunct>

      {<space> OR <space> <disjunct>}

      <disjunct> ::= <term>

      {<space> <term>}

      <term> ::= ['-'] <simple term>
      <simple term> ::= <word> | <phrase>
      <word> ::= <non space char>

      {<non space char>}

      <phrase> ::= '"' <word>

      {<space> <word>}

      '"'
      <space> ::= <space char>

      {<space char>}

      <non space char> ::= <char> - <space char> /* Any Char except SpaceChar */
      <space char> ::= ' '
      <char> ::= /* Any character */

        1. Proposal 1 Semantics

      All proposal 0 semantics, plus

      • Terms separated by whitespace are implicitly "ANDed"
      • Terms separated by "OR" are "ORed"
      • "AND" has higher precedence than "OR"
      • Within a word, each double quote (") must also be escaped by a preceding "\" (backslash)
      Show
      As defined by JSR-283, but modified to the style of the CMIS SQL grammar. Note: <space> & <non space char> definitions taken directly from JSR-283, but consideration of whitespace is probably needed. <text search expression> ::= <disjunct> {<space> OR <space> <disjunct>} <disjunct> ::= <term> {<space> <term>} <term> ::= ['-'] <simple term> <simple term> ::= <word> | <phrase> <word> ::= <non space char> {<non space char>} <phrase> ::= '"' <word> {<space> <word>} '"' <space> ::= <space char> {<space char>} <non space char> ::= <char> - <space char> /* Any Char except SpaceChar */ <space char> ::= ' ' <char> ::= /* Any character */ Proposal 1 Semantics All proposal 0 semantics, plus Terms separated by whitespace are implicitly "ANDed" Terms separated by "OR" are "ORed" "AND" has higher precedence than "OR" Within a word, each double quote (") must also be escaped by a preceding "\" (backslash)
    • Hide

      Proposal 1 in doc in tc

      Show
      Proposal 1 in doc in tc

      The text search expression is defined as a <character string literal> (as defined by SQL-92). However, the syntax and semantics of the full text search expression are repo specific.

      I remember there was some resistance to defining a 'lowest common denominator' full text search language, but I don't remember why.

      Given that we define SQL, and that query is a key use case, I think there's value in a deeper FTS definition.

      As a starting point, JCR provides minimal definition. I'm not sure we would need to much further than that to start with.

            Assignee:
            Ethan Gur-esh (Inactive)
            Reporter:
            David Caruana (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

              Created:
              Updated:
              Resolved: