[
https://issues.apache.org/jira/browse/LUCENE-4644?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13609906#comment-13609906
]
David Smiley commented on LUCENE-4644:
--------------------------------------
It just dawned upon me that this limitation of Within finding false-positives
could be remedied right now, without waiting for the faster approach in
LUCENE-4869. I see 3 configurable options:
# As indicated in earlier discussion on this issue on the first comment, it can
visit all regions outside the query shape and mark those docs for exclusion in
the final results. Yes this would be slow but it would work, and it would be
particularly easy to implement as my existing code would only need a small
modification. So in summary, this approach can be addressed now, it's slow, it
works.
# Allow a configurable buffer where the user wants Within, but is willing to
disregard any regions that might be outside of that buffer that would create a
false-positive match. Is this useful? It's easy to add to the existing code
which is already doing shape buffering.
# The approach in LUCENE-4869: filter false-positives out using a
representative point from each disjoint piece of an indexed shape that is saved
at index time. This will be fast but will require memory and it'll take some
time to get there based on where the APIs are at now.
> Implement spatial WITHIN query for RecursivePrefixTree
> ------------------------------------------------------
>
> Key: LUCENE-4644
> URL: https://issues.apache.org/jira/browse/LUCENE-4644
> Project: Lucene - Core
> Issue Type: New Feature
> Components: modules/spatial
> Reporter: David Smiley
> Assignee: David Smiley
> Attachments:
> LUCENE-4644_Spatial_Within_predicate_for_RecursivePrefixTree.patch
>
>
--
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]
