You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "David Smiley (JIRA)" <ji...@apache.org> on 2017/11/16 18:43:00 UTC
[jira] [Updated] (LUCENE-8050) PerFieldDocValuesFormat's merge
should not grab field DVF if DocValuesType.NONE
[ https://issues.apache.org/jira/browse/LUCENE-8050?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
David Smiley updated LUCENE-8050:
---------------------------------
Attachment: LUCENE_8050.patch
Here's a patch. I can write a test.
WDYT [~mikemccand] -- you shepherded in LUCENE-7456.
> PerFieldDocValuesFormat's merge should not grab field DVF if DocValuesType.NONE
> -------------------------------------------------------------------------------
>
> Key: LUCENE-8050
> URL: https://issues.apache.org/jira/browse/LUCENE-8050
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/index
> Affects Versions: 6.3
> Reporter: David Smiley
> Assignee: David Smiley
> Attachments: LUCENE_8050.patch
>
>
> Since LUCENE-7456 (Lucene 6.3), PerFieldDocValuesFormat delegates the merge to the actual field DVF's merge. Great, but unfortunately it will call {{getDocValuesFormatForField}} on all fields (in FieldInfos) even those that have no DocValues (DocValuesType.NONE). It won't ultimately actually write anything to those DVFs but there may be some overhead and furthermore it's now more awkward to write a subclass of PFDVF that deliberately throws an exception from {{getDocValuesFormatForField}} for some fields.
> AFAICT this appears to be a non-issue for PerFieldPostingsFormat's merge because it's use of MultiFields filters out IndexOptions.NONE
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org