[
https://issues.apache.org/jira/browse/LUCENE-2617?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12901896#action_12901896
]
Michael McCandless commented on LUCENE-2617:
--------------------------------------------
Amazingly, I just hit a random test failure (seed = 5148889059941714836L
reproduces it):
{noformat}
[junit] Testsuite: org.apache.lucene.search.TestMultiSearcherRanking
[junit] Testcase:
testTwoTermQuery(org.apache.lucene.search.TestMultiSearcherRanking): FAILED
[junit] expected:<0.48314467> but was:<0.60746753>
[junit] junit.framework.AssertionFailedError: expected:<0.48314467> but
was:<0.60746753>
[junit] at
org.apache.lucene.search.TestMultiSearcherRanking.checkQuery(TestMultiSearcherRanking.java:102)
[junit] at
org.apache.lucene.search.TestMultiSearcherRanking.testTwoTermQuery(TestMultiSearcherRanking.java:47)
[junit] at
org.apache.lucene.util.LuceneTestCase.runBare(LuceneTestCase.java:380)
[junit] at
org.apache.lucene.util.LuceneTestCase.run(LuceneTestCase.java:372)
{noformat}
The failure is in fact caused by this issue! (The patch fixes it). The
RandomIndexWriter in this test w/ the above seed creates a multi-segment index,
and then tests on a 2-term BQ. Some segments did/didn't have both of the terms
and this causes the score difference.
So... I think we really should fix this, for consistency. Otherwise, how your
docs fall into segments can alter your scores, which is awful.
> coord should still apply to missing terms/clauses
> -------------------------------------------------
>
> Key: LUCENE-2617
> URL: https://issues.apache.org/jira/browse/LUCENE-2617
> Project: Lucene - Java
> Issue Type: Bug
> Affects Versions: 4.0
> Reporter: Yonik Seeley
> Attachments: LUCENE-2617.patch
>
>
> Missing terms in a boolean query "disappear" (i.e. they don't even affect the
> coord factor).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
