[
https://issues.apache.org/jira/browse/LUCENE-5476?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Shai Erera updated LUCENE-5476:
-------------------------------
Attachment: LUCENE-5476.patch
I reviewed the patch more closely before I commit:
* Modified few javadocs
* Removed needsSampling() since we don't offer an extension any access to e.g.
totalHits. We can add it when there's demand.
* Fixed a bug in how carryOver was implemented -- replaced by two members
{{leftoverBin}} and {{leftoverIndex}}. So now if {{leftoverBin != -1}} we know
to skip the first such documents in the next segment and depending on
{{leftoverIndex}}, whether we need to sample any of them. Before that, we
didn't really skip over the leftover docs in the bin, but started to count a
new bin.
* Added a CHANGES entry.
I reviewed the test - would be good if we can write a unit test which
specifically matches only few documents in one segment compared to the rest. I
will look into it perhaps later.
I think it's ready, but if anyone wants to give createSample() another look, to
make sure this time leftover works well, I won't commit it by tomorrow anyway.
> Facet sampling
> --------------
>
> Key: LUCENE-5476
> URL: https://issues.apache.org/jira/browse/LUCENE-5476
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Rob Audenaerde
> Attachments: LUCENE-5476.patch, LUCENE-5476.patch, LUCENE-5476.patch,
> LUCENE-5476.patch, LUCENE-5476.patch, LUCENE-5476.patch, LUCENE-5476.patch,
> LUCENE-5476.patch, LUCENE-5476.patch, LUCENE-5476.patch, LUCENE-5476.patch,
> SamplingComparison_SamplingFacetsCollector.java, SamplingFacetsCollector.java
>
>
> With LUCENE-5339 facet sampling disappeared.
> When trying to display facet counts on large datasets (>10M documents)
> counting facets is rather expensive, as all the hits are collected and
> processed.
> Sampling greatly reduced this and thus provided a nice speedup. Could it be
> brought back?
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
