You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Simon Willnauer (JIRA)" <ji...@apache.org> on 2014/07/05 13:42:33 UTC

[jira] [Commented] (LUCENE-5784) CommonTermsQuery HighFreq MUST not applied if lowFreq terms

    [ https://issues.apache.org/jira/browse/LUCENE-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14052843#comment-14052843 ] 

Simon Willnauer commented on LUCENE-5784:
-----------------------------------------

actually the idea here was that the operator for high freq is only to control when / if the score for that query should be taken into account. I personally think we should never make the high freq required otherwise we loose all the performance gains here possibly? Maybe it's actually not that bad though since we have the low freq terms driving the matching... I guess we could say that we let the minimumShouldMatch on the highfreq query control when and if the query is used for scoring and make the terms there always SHOULD clauses and use the highFreq occur as the occur used on the toplevel query?

> CommonTermsQuery HighFreq MUST not applied if lowFreq terms
> -----------------------------------------------------------
>
>                 Key: LUCENE-5784
>                 URL: https://issues.apache.org/jira/browse/LUCENE-5784
>             Project: Lucene - Core
>          Issue Type: Bug
>          Components: core/query/scoring
>    Affects Versions: 4.8.1
>            Reporter: Clinton Gormley
>            Priority: Minor
>         Attachments: common_terms.patch
>
>
> When a CommonTermsQuery has high and low frequency terms,  the highFreq terms Boolean query is always added as a SHOULD clause, even if highFreqOccur is set to MUST:
>     new CommonTermsQuery(Occur.MUST, Occur.MUST,0.1);
> My patch sets the top level Boolean query's minimum should match to 1 to ensure that the SHOULD clause must match.  Not sure if this is the correct approach, or if it should just add the highFreq query as a MUST clause instead?



--
This message was sent by Atlassian JIRA
(v6.2#6252)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org