Re: svn commit: r1646469 - in /jackrabbit/oak/trunk/oak-solr-core/src: main/java/org/apache/jackrabbit/oak/plugins/index/solr/index/ main/java/org/apache/jackrabbit/oak/plugins/index/solr/osgi/ main/java/org/apache/jackrabbit/oak/plugins/index/solr/query/ ...

[Chetan Mehrotra <ch...@gmail.com>]

Hi Tommaso,

On Thu, Dec 18, 2014 at 8:10 PM,  <to...@apache.org> wrote:
> Log:
> OAK-2168 - support qt aggregation (wip)

I think query time aggregation over a remote index like Solr would not
be able to perform well. Specially for multi term queries. So it might
be better if we only add support for index time aggregation.

Chetan Mehrotra

Re: svn commit: r1646469 - in /jackrabbit/oak/trunk/oak-solr-core/src: main/java/org/apache/jackrabbit/oak/plugins/index/solr/index/ main/java/org/apache/jackrabbit/oak/plugins/index/solr/osgi/ main/java/org/apache/jackrabbit/oak/plugins/index/solr/query/ ...

[Tommaso Teofili <to...@gmail.com>]

Hi Chetan,

I agree index time aggregation will be probably faster but I'd like the
Solr index to be able to do both so that it's on par with the Lucene index
features. On a side note the mentioned commit would only work pre-OAK-2119
so the next step will be (I'm currently working on it) to properly move to
implement the AdvanceQueryIndex.

On a related note it'd be probably good to add some benchmarks on query
time / index time aggregation performances for Lucene (and for Solr once
available) to have a proper evidence of what we expect.

Regards,
Tommaso

2014-12-19 6:33 GMT+01:00 Chetan Mehrotra <ch...@gmail.com>:
>
> Hi Tommaso,
>
> On Thu, Dec 18, 2014 at 8:10 PM,  <to...@apache.org> wrote:
> > Log:
> > OAK-2168 - support qt aggregation (wip)
>
> I think query time aggregation over a remote index like Solr would not
> be able to perform well. Specially for multi term queries. So it might
> be better if we only add support for index time aggregation.
>
> Chetan Mehrotra
>