JCR-Query: rep:excerpt is empty from subnodes aggregated in lucene searchindex

["Dr. Hans-Peter Störr" <yu...@techno.ms>]

Hi!

We are using Jackrabbit JAR (version 2.14.0) within the Sling framework as a content repository. For the implementation of a full text search we switch on aggregation in the lucene search index such that the subnodes of nodes of type cpp:PageContext are aggregated into that node (compare https://jackrabbit.apache.org/oak/docs/query/lucene.html#aggregation):

aggregates
    cpp:Page
        include0
            path="*"
        include1
            path="*/*"
and so forth.

A JCR-SQL2 fulltext search like

SELECT excerpt(n) FROM [cpp:PageContent] AS n 
WHERE ISDESCENDANTNODE(n, '/content') AND (CONTAINS(n.*,'searchterm' ) )  

finds nodes of type cpp:PageContent that have the searchterm in their attributes and nodes that have the searchterm in the attributes of their subnodes, as required by our application. My problem is however, that the excerpt (https://wiki.apache.org/jackrabbit/ExcerptProvider) contains something if the searchterm is found in the properties of the returned node, but nothing when the searchterm is only found in the properties of the subnodes that have been aggregated.

Is there anything I can do to retrieve the excerpt for the subnodes without traversing the whole (potentially sizeable) hierarchy of each of the several returned search results? Is there any special syntax I could use, as http://jackrabbit.510166.n4.nabble.com/How-to-get-excerpt-for-child-node-td514886.html suggests? (Unfortunately the solution given there does not work for me, since I don't know at which subnodes the terms are found, and if there are several searchterms these might even be found on different subnodes.) Thanks so much!

Best regards,

Hans-Peter