[
https://issues.apache.org/jira/browse/SOLR-106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12465254
]
Yonik Seeley commented on SOLR-106:
-----------------------------------
Thanks for the info JJ... didn't see your update untill after I committed this
(I'm running a bit behind all the solr traffic :-)
> Case for Facet Count Caching: Paging through the hitlist
Hmmm, yes that would be good for a more stateless client. Even more efficient
would be to recognize in the client that since you are only changing a page in
the hitlist, the facets won't change (and hence don't re-query).
> It occurs to me that facet.limit should NOT do double-duty for paging:
Or, it should *only* be used for paging, specifying the number to be returned.
The BoundedTreeSet size and caching are an implementation detail and shouldn't
be in the API unless really necessary. If it matters in the future, we could
add a hint specifying how much "extra" should be computed.
> Case for pulling response generation out of getFieldCacheCounts and
> getFacetTermEnumCounts
Sure, makes sense. Don't view the current facet code as "done"... I have a
*lot* of little ideas on how to make it better, esp for cases like faceting on
author.
> TermFreqVectors
Regarding this, do you have any performance data on it... my assumption was
that it would be too slow for a large number of hits. Perhaps still a good
option to have if the number of hits are small and the fieldcache isn't an
option though.
> Just had an idea: It would be even nicer if the counting logic could be
> passed some object,
Yup, separating those things was on my todo list.
> new facet params: facet.sort, facet.mincount, facet.offset
> ----------------------------------------------------------
>
> Key: SOLR-106
> URL: https://issues.apache.org/jira/browse/SOLR-106
> Project: Solr
> Issue Type: Improvement
> Components: search
> Reporter: Yonik Seeley
> Attachments: facet_params.patch
>
>
> a couple of new facet params:
> facet lists become pageable with facet.offset, facet.limit (idea from Erik)
> facet.sort explicitly specifies sort order (true for count descending, false
> for natural index order)
> facet.mincount: minimum count for facets included in response (idea from JJ,
> deprecate zeros)
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
