Details

    • Type: Bug Bug
    • Status: Closed
    • Priority: Medium Medium
    • Resolution: Obsolete
    • Affects Version/s: None
    • Fix Version/s: Future
    • Labels:
      None
    • Environment:

      Operating System: Ubuntu 9.10
      PHP Version: '5.2.10-2ubuntu6.3 with Suhosin-Patch 0.9.7'
      Database and version:'mysql 5.1.37'
      Browser (and version):
      installations: ezpublish-4.2.0-full-gpl and ezpublish 4.0.3 with ezfind 2.1.0

      Description

      Hello.
      I created an object that includes German umlauts (ä,ü or ö) both in the title and in the text.
      If I search for the text in combination with a wildcard, I receive no result.
      If I replace the german umlaut by "a, u or o" I receive a result.

      for Example:

      Object title: "Erörterung"

      Searching for: "erör *"----no search result
      Searching for: "eror *"----object is found
      Searching for: "erörterung"----object is found

        Issue Links

          Activity

          Hide
          Paul Borgermans added a comment -

          Yes, this is a known ssue in Solr/Lucene: in case of wildcard and also range searches, the analyzers used for indexing are not called (becomes also case sensitive, while normally all tokens are lowercased)

          Out of the box ez find does normalisation, so ö becomes oe in the index

          We'll need some client side (PHP) manipulation, because is seems the issue won't be resolved in Solr/Lucene

          Show
          Paul Borgermans added a comment - Yes, this is a known ssue in Solr/Lucene: in case of wildcard and also range searches, the analyzers used for indexing are not called (becomes also case sensitive, while normally all tokens are lowercased) Out of the box ez find does normalisation, so ö becomes oe in the index We'll need some client side (PHP) manipulation, because is seems the issue won't be resolved in Solr/Lucene
          Hide
          Richard Bayet added a comment -

          Should be linked to #014906 : "Wildcard-Queries are not lowercased and therefore lead to no results".

          Show
          Richard Bayet added a comment - Should be linked to #014906 : "Wildcard-Queries are not lowercased and therefore lead to no results".
          Hide
          (inactive) Gunnstein Lye added a comment -

          This is caused by related issue:
          http://issues.ez.no/11673
          which adds ISOLatin1AccentFilterFactory to schema.xml. This filter doesn't run on wildcard queries, causing them to fail.

          The solution is to revert #11673, then restart Solr, and reindex all affected objects.
          The tradeoff for this fix is that the search will not be robust against spelling mistakes such as "e" being used rather than "é" (accented e).

          Show
          (inactive) Gunnstein Lye added a comment - This is caused by related issue: http://issues.ez.no/11673 which adds ISOLatin1AccentFilterFactory to schema.xml. This filter doesn't run on wildcard queries, causing them to fail. The solution is to revert #11673, then restart Solr, and reindex all affected objects. The tradeoff for this fix is that the search will not be robust against spelling mistakes such as "e" being used rather than "é" (accented e).
          Hide
          (inactive) Gunnstein Lye added a comment -

          In reply to comment #063556
          Patch for reverting #11673 on ezfind 2.3.issue-8729-ezfind-2.3.diff

          Show
          (inactive) Gunnstein Lye added a comment - In reply to comment #063556 Patch for reverting #11673 on ezfind 2.3. issue-8729-ezfind-2.3.diff
          Hide
          ezrobot added a comment -

          This issue has been automatically closed due to the lack of activity over a long period of time. It is very likely that it is obsolete, but if you think it is still valid, do not hesitate to reopen it and mention why.

          Show
          ezrobot added a comment - This issue has been automatically closed due to the lack of activity over a long period of time. It is very likely that it is obsolete, but if you think it is still valid, do not hesitate to reopen it and mention why.

            People

            • Assignee:
              Paul Borgermans
              Reporter:
              Miguel Seuster
            • Votes:
              0 Vote for this issue
              Watchers:
              0 Start watching this issue

              Dates

              • Created:
                Updated: