[
https://issues.apache.org/jira/browse/LUCENE-4312?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16880828#comment-16880828
]
Robert Muir commented on LUCENE-4312:
-------------------------------------
I don't think chicken and the egg description works well as an argument for
something to add to the index. We should have a high bar in order to do that,
because once something gets added its basically impossible to remove.
My earlier suggestion (payloads) was based on the fact that we are talking
about corner-cases as far as search improvements, at a heavy complexity cost.
Maybe we could first address the search side with payload-based queries (maybe
in sandbox, similar to what you already developed?) to try to address
[~jpountz] concerns about scalability before actually optimizing it further by
encoding in the index?
This way it wouldn't have to be all solved at once.
> Index format to store position length per position
> --------------------------------------------------
>
> Key: LUCENE-4312
> URL: https://issues.apache.org/jira/browse/LUCENE-4312
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/codecs
> Affects Versions: 6.0
> Reporter: Gang Luo
> Priority: Minor
> Labels: Suggestion
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> Mike Mccandless said:TokenStreams are actually graphs.
> Indexer ignores PositionLengthAttribute.Need change the index format (and
> Codec APIs) to store an additional int position length per position.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
