[
https://issues.apache.org/jira/browse/LUCENE-3196?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13048787#comment-13048787
]
Uwe Schindler commented on LUCENE-3196:
---------------------------------------
Looks fine, using this approach, separate norms impl can hopefully go away
quite fast *g*
For the PreFlex codec I even have an idea for the codec and backwards
compatibility: The old norms file could be exposed as standard DocValues field
by PreFlex codec. The r/w StandardCodec would never write separate norms files,
instead simply write docvalues using this 1 byte approach (of course
configureable to have e.g. read float norms, and other additional BM25
statistics or whatever).
Just ideas, Uwe
> Optimize FixedStraightBytes for bytes size == 1
> -----------------------------------------------
>
> Key: LUCENE-3196
> URL: https://issues.apache.org/jira/browse/LUCENE-3196
> Project: Lucene - Java
> Issue Type: Improvement
> Components: core/index, core/search
> Affects Versions: 4.0
> Reporter: Simon Willnauer
> Assignee: Simon Willnauer
> Priority: Minor
> Fix For: 4.0
>
> Attachments: LUCENE-3196.patch
>
>
> Currently we read all the bytes in a PagedBytes instance wich is unneeded for
> single byte values like norms. For fast access this should simply be a
> straight array.
--
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]
