[
https://issues.apache.org/jira/browse/LUCENE-3507?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13125153#comment-13125153
]
Michael McCandless commented on LUCENE-3507:
--------------------------------------------
Wow very impressive!
So now we don't load any of the byte[] nor ord data for the subs
during merging (we use the new DirectSource); we only allocate the ord
re-mapping per segment, and docToOrd array for the merged segment.
Very nice!
Patch looks great. Only a couple tiny things:
* In SortedBytesMergeUtils.createOrdMapping, for the deletes case, I
think you can move the {{int ord = currentSlice.source.ord(j)}} inside
the if? Ie we don't need to pull the ord unless doc is still live.
* Typo in SortedBytesMergeUtils: mapps -> maps
> Improve Memory Consumption for merging DocValues SortedBytes variants
> ---------------------------------------------------------------------
>
> Key: LUCENE-3507
> URL: https://issues.apache.org/jira/browse/LUCENE-3507
> Project: Lucene - Java
> Issue Type: Improvement
> Components: core/index
> Affects Versions: 4.0
> Reporter: Simon Willnauer
> Assignee: Simon Willnauer
> Fix For: 4.0
>
> Attachments: LUCENE-3507.patch
>
>
> Currently SortedBytes are loaded into memory during merge which could be a
> potential trap. Instead of loading them into Heap memory we can merge those
> sorted values with much smaller memory and without loading all values into
> ram.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
