lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yoni Amir (JIRA)" <>
Subject [jira] [Created] (SOLR-3919) Problem with hl.mergeContinuous - sniipt is too long but only one hit
Date Sun, 07 Oct 2012 12:38:02 GMT
Yoni Amir created SOLR-3919:

             Summary: Problem with hl.mergeContinuous - sniipt is too long but only one hit
                 Key: SOLR-3919
             Project: Solr
          Issue Type: Bug
          Components: highlighter
    Affects Versions: 4.0-BETA
         Environment: win xp, java 7
            Reporter: Yoni Amir

I am using a configuration roughly as follows:

       <bool name="hl.usePhraseHighlighter">true</bool>
       <bool name="hl.highlightMultiTerm">true</bool>
       <int name="hl.snippets">4</int>
       <bool name="hl.mergeContiguous">true</bool>
I found a strange case as follows:

I have only 1 hit in the field, so I expect that no merging will take place. However, solr
returns a highlight snippet of length 400. In effect, it calculates the hl.snippet size times
100. With hl.snippet=10, the length is 1000. I think it is because hl.fragsize is 100 by default.

This occurs only if the hit is sufficiently close to the beginning of the field's text - that
is, if it is between the 300th and 400th character. In that change, the first four 100-block
size snippets are merged as if all of them have a hit.

This behavior is wrong, and if there is only one hit, I don't expect solr to merge anything
for me.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message