[jira] [Created] (OAK-944) Full-text index aggregation

["Thomas Mueller (JIRA)" <ji...@apache.org>]

Thomas Mueller created OAK-944:
----------------------------------

             Summary: Full-text index aggregation
                 Key: OAK-944
                 URL: https://issues.apache.org/jira/browse/OAK-944
             Project: Jackrabbit Oak
          Issue Type: Bug
          Components: query
            Reporter: Thomas Mueller


Jackrabbit 2.x supports index aggregates as documented here:
http://wiki.apache.org/jackrabbit/IndexingConfiguration#Index_Aggregates

Jackrabbit Oak should support this feature as well.

There are multiple way to implement it:

* Aggregate the data when indexing. This is what Jackrabbit 2.x does. The disadvantage is that if filtering is needed (see also JCR-3611), then this needs to be done at indexing time; and filtering at query time might be hard to achieve.

* Aggregate the data when querying. In this case, filtering can be implemented at query execution time.

Alex, Marcel and me yesterday briefly discussed the options and came to the conclusion that we try use the second approach for Oak, that is, aggregate the data when querying. One way to achieve that is by adding a new FulltextQueryIndex which does the filtering and aggregation. This index internally uses a FulltextQueryIndex similar to how the NodeTypeIndex uses a PropertyIndex. This "aggregating full-text index" can then be used for any kind of full-text index (Lucene, Solr,...). 

To allow more flexibility in aggregate filtering, we will extend the Filter interface to allow reading from the repository (using the Tree interface).


--
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