lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "David Smiley (JIRA)" <>
Subject [jira] [Commented] (SOLR-3229) TermVectorComponent does not return terms in distributed search
Date Tue, 15 Dec 2015 19:04:47 GMT


David Smiley commented on SOLR-3229:

[~hossman] HighlightComponent, DebugComponent, and TermVectorComponent have a very similar
bit of code in their finishStage() method. HighlightComponent & DebugComponent's version
was recently found to be buggy -- SOLR-8060 and SOLR-8059.  The Highlight side was recently
fixed and I'm about to do the same for Debug side.  But I'd like to refactor out some common
lines of code between all 3 to ease maintenance.  However the TV side has this odd bit where
it if it can't lookup the shard doc by it's unique key, that it adds this to the response
any way (~line 458).  _I would rather we remove this; I think it's not something we should
support.  UniqueKey should be required for distributed-search to get TV info back_.  The code
that's here now incorrectly assumes that if it was unable to lookup the key in the resultIds
that it's because the schema has no uniqueKey.  But another reason is just that it's a distrib.singlePass
distributed search (related to the 2 bugs I'm looking at in Highlight & Debug components).
 Do you support my recommendation?

> TermVectorComponent does not return terms in distributed search
> ---------------------------------------------------------------
>                 Key: SOLR-3229
>                 URL:
>             Project: Solr
>          Issue Type: Bug
>          Components: SearchComponents - other
>    Affects Versions: 4.0-ALPHA
>         Environment: Ubuntu 11.10, openjdk-6
>            Reporter: Hang Xie
>            Assignee: Hoss Man
>              Labels: patch
>             Fix For: 4.0, Trunk
>         Attachments: SOLR-3229.patch, TermVectorComponent.patch
> TermVectorComponent does not return terms in distributed search, the distributedProcess()
incorrectly uses Solr Unique Key to do subrequests, while process() expects Lucene document
ids. Also, parameters are transferred in different format thus making distributed search returns
no result.

This message was sent by Atlassian JIRA

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

View raw message