[
https://issues.apache.org/jira/browse/LUCENE-3146?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13041252#comment-13041252
]
Michael McCandless commented on LUCENE-3146:
--------------------------------------------
It's true that's a risk but... omitNorms (and omitTF) are effectively today a
"global schema" because on merge Lucene cannot preserve different settings
here, unlike, eg, term vectors or stored fields or the upcoming doc values,
where every document is free to have its own "schema".
The fact that apps now see different behavior depending on when Lucene happened
to merge segments is awful.
Catching/enforcing this global schema up front, now made possible because we
have the fnx file, is cleaner because merge selection would no longer be
"visible" in the API.
> IndexReader.setNorms is no op if one of the field instances omits norms
> -----------------------------------------------------------------------
>
> Key: LUCENE-3146
> URL: https://issues.apache.org/jira/browse/LUCENE-3146
> Project: Lucene - Java
> Issue Type: Bug
> Components: core/search
> Reporter: Shai Erera
> Fix For: 4.0
>
> Attachments: LUCENE-3146.patch
>
>
> If I add two documents to an index w/ same field, and one of them omit norms,
> then IndexReader.setNorms is no-op. I'll attach a patch w/ test case
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
