[
https://issues.apache.org/jira/browse/LUCENE-6163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14264677#comment-14264677
]
Tomoko Uchida commented on LUCENE-6163:
---------------------------------------
Thanks for the response.
> I don't think the optimization is safe if any tokenfilter modifies the tokens
> after ngram, it will give the wrong results.
I understand your indication. Your point is about my patch, not about
NGramPhraseQuery itself?
If my approach is bad, another way, as noted in the past comment of SOLR-3055,
could still be possible.
any suggestions will be appreciated.
> Use NGramPhraseQuery via QueryBuilder
> -------------------------------------
>
> Key: LUCENE-6163
> URL: https://issues.apache.org/jira/browse/LUCENE-6163
> Project: Lucene - Core
> Issue Type: New Feature
> Components: core/other, modules/analysis
> Affects Versions: 5.0
> Reporter: Tomoko Uchida
> Priority: Minor
> Attachments: LUCENE-6163.patch
>
>
> This is derived from SOLR-3055.
> To gain performance improvement of LUCENE-3426, QueryBuilder should generate
> NGramPhraseQuery to optimize n-gram based phrase query.
> (especially good for CJK languages, but the benefit is not limited to those.)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
