Solr
  1. Solr
  2. SOLR-4612

Admin UI - Analysis Screen contains empty table-columns

    Details

    • Type: Bug Bug
    • Status: Closed
    • Priority: Minor Minor
    • Resolution: Fixed
    • Affects Version/s: 4.1
    • Fix Version/s: 4.7, 6.0
    • Component/s: web gui
    • Labels:
      None

      Description

      While checking the links from the Tutorial, i discovered (at least) one which contains multiple empty table columns in the generated result-table

      http://localhost:8983/solr/#/collection1/analysis?analysis.fieldvalue=%EF%BD%B6%EF%BE%80%EF%BD%B6%EF%BE%85&analysis.query=%E3%82%AB%E3%82%BF%E3%82%AB%E3%83%8A&analysis.fieldtype=text_cjk&verbose_output=0

      ST カタカナ  
      CJKWF カタカナ  
      LCF カタカナ  
      CJKBF カタ, タカ, カナ  

      (because the jira-markup allows not colspan, used a comma to visualize the three columns of the last row)

      Not sure where they are coming from ..

      1. SOLR-4612.patch
        2 kB
        Stefan Matheis (steffkes)

        Issue Links

          Activity

          Hide
          Stefan Matheis (steffkes) added a comment -

          After a bit digging into the code, the problem appears to be: the CJKBigramFilter splits the provided into three parts (the code therefore raises the column-count to three), but since they belong all to the first word, the table later on uses html's colspan to make the relation clear. results in a table with three columns, where only the first one is used - because the first analyzer only gives one word.

          perhaps a combination with positionHistory would work, since that is later on used for the colspan magic, which i already mentioned ..

          Show
          Stefan Matheis (steffkes) added a comment - After a bit digging into the code, the problem appears to be: the CJKBigramFilter splits the provided into three parts (the code therefore raises the column-count to three), but since they belong all to the first word, the table later on uses html's colspan to make the relation clear. results in a table with three columns, where only the first one is used - because the first analyzer only gives one word. perhaps a combination with positionHistory would work, since that is later on used for the colspan magic, which i already mentioned ..
          Hide
          Stefan Matheis (steffkes) added a comment -

          looks like it could be done much easier than i initially thought.

          the first item in the chain defines the upper limit of possible positions - and the count never gets higher than that.

          so to say, the position itself does, but the positionHistory never does - and this is what the UI relies on. simply example would be "TvTv" with a WDF, which gets split into two tokens but on the same position (read as positionHistory)

          works at least for the given sample and a few others which i've tested randomly. needs a bit more testing to make sure that nothing gets broken w/ this patch.

          Show
          Stefan Matheis (steffkes) added a comment - looks like it could be done much easier than i initially thought. the first item in the chain defines the upper limit of possible positions - and the count never gets higher than that. so to say, the position itself does, but the positionHistory never does - and this is what the UI relies on. simply example would be "TvTv" with a WDF , which gets split into two tokens but on the same position (read as positionHistory ) works at least for the given sample and a few others which i've tested randomly. needs a bit more testing to make sure that nothing gets broken w/ this patch.
          Hide
          ASF subversion and git services added a comment -

          Commit 1542547 from Stefan Matheis (steffkes) in branch 'dev/trunk'
          [ https://svn.apache.org/r1542547 ]

          SOLR-4612: Admin UI - Analysis Screen contains empty table-columns

          Show
          ASF subversion and git services added a comment - Commit 1542547 from Stefan Matheis (steffkes) in branch 'dev/trunk' [ https://svn.apache.org/r1542547 ] SOLR-4612 : Admin UI - Analysis Screen contains empty table-columns
          Hide
          ASF subversion and git services added a comment -

          Commit 1542548 from Stefan Matheis (steffkes) in branch 'dev/branches/branch_4x'
          [ https://svn.apache.org/r1542548 ]

          SOLR-4612: Admin UI - Analysis Screen contains empty table-columns (merge r1542547)

          Show
          ASF subversion and git services added a comment - Commit 1542548 from Stefan Matheis (steffkes) in branch 'dev/branches/branch_4x' [ https://svn.apache.org/r1542548 ] SOLR-4612 : Admin UI - Analysis Screen contains empty table-columns (merge r1542547)
          Hide
          Hossein Taghi-Zadeh added a comment - - edited

          I has a fieldType which used the two separate PatternReplaceCharFilterFactory.
          for example
          <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="[\u200E|\u200C|\u200D|\u200F]\s" replacement=" " />
          <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="(?<Unable to render embedded object: File ((\d) not found.(\d|\w))" replacement=" " />
          <tokenizer class="solr.StandardTokenizerFactory"/>

          And in analysis screen, after PRCF filter any result was not shown!!!

          Show
          Hossein Taghi-Zadeh added a comment - - edited I has a fieldType which used the two separate PatternReplaceCharFilterFactory. for example <charFilter class="solr.PatternReplaceCharFilterFactory" pattern=" [\u200E|\u200C|\u200D|\u200F] \s" replacement=" " /> <charFilter class="solr.PatternReplaceCharFilterFactory" pattern="(?< Unable to render embedded object: File ((\d) not found. (\d|\w))" replacement=" " /> <tokenizer class="solr.StandardTokenizerFactory"/> And in analysis screen, after PRCF filter any result was not shown!!!
          Hide
          Shawn Heisey added a comment -

          Hossein Taghi-Zadeh, I assume that you are using the just-released 4.7.0, or perhaps a later dev version. SOLR-5800 has been filed for the problem you are experiencing.

          Show
          Shawn Heisey added a comment - Hossein Taghi-Zadeh , I assume that you are using the just-released 4.7.0, or perhaps a later dev version. SOLR-5800 has been filed for the problem you are experiencing.

            People

            • Assignee:
              Stefan Matheis (steffkes)
              Reporter:
              Stefan Matheis (steffkes)
            • Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development