[
https://issues.apache.org/jira/browse/LUCENE-5293?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13800143#comment-13800143
]
Paul Elschot commented on LUCENE-5293:
--------------------------------------
After some more thought on this I think using the WA8 docidset as in the patch
is the best solution for now, because I think that gives the best building time
for the expected cases.
I might add an EliasFanoEncoder constructor with only an upperBound argument
for this case.
This would leave some room for adding more values (as in ArrayUtil.grow) and it
would reorganize the encoded sequence to always use the latest number of
values. Reorganizing the encoded sequence would be needed when the number of
bits for encoding the lower values changes, and this is
floor(log2(upperBound/numValues)) but never negative.
(In a docidset for filtering the upperBound is normally the segment size, and
the values are the doc ids.)
> Also use EliasFanoDocIdSet in CachingWrapperFilter
> --------------------------------------------------
>
> Key: LUCENE-5293
> URL: https://issues.apache.org/jira/browse/LUCENE-5293
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/search
> Reporter: Paul Elschot
> Priority: Minor
> Attachments: LUCENE-5293.patch, LUCENE-5293.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.1#6144)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
