[
https://issues.apache.org/jira/browse/LUCENE-4752?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13594638#comment-13594638
]
Adrien Grand commented on LUCENE-4752:
--------------------------------------
bq. I took a look at MP API, and maybe if we change OneMerge from holding a
List<SegmentReader> to List<AtomicReader>, we could write an MP which sorts
segments together by opening a SortingAtomicReader over the segments that were
picked for merge?
Making the sorting stuff part of MergePolicy makes sense. However, I think that
the (package-private) List<SegmentReader> in MergePolicy is only used to track
the list of segment readers being used while merging (this reference is only
used in IndexWriter). What MP actually manipulates is a list of
SegmentInfoPerCommit, it is possible that no reader is open for a segment when
MergePolicy picks it, and Lucene should not force a reader to be open until the
merge actually starts. So maybe we should have an additional method in
MergePolicy (or OneMerge for finer-grained control?) to tell IndexWriter how to
view a list of segment readers? (either sequentially as today or a sorted view
as suggested in this issue description).
> Merge segments to sort them
> ---------------------------
>
> Key: LUCENE-4752
> URL: https://issues.apache.org/jira/browse/LUCENE-4752
> Project: Lucene - Core
> Issue Type: New Feature
> Components: core/index
> Reporter: David Smiley
> Assignee: Adrien Grand
>
> It would be awesome if Lucene could write the documents out in a segment
> based on a configurable order. This of course applies to merging segments
> to. The benefit is increased locality on disk of documents that are likely to
> be accessed together. This often applies to documents near each other in
> time, but also spatially.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
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]
