Dear all,
I am running a grouped query including facets in my Junit Test cases
against a Solr 4.2.1 Embedded Server. When faceting the groups, I want the
counts to reflect the number of groups, not the number of documents. But
when I enable "&group.facet=true" on the query, the test fails with the
following message:
*** BEGIN testSearchByQuery(com.test.InsaneFieldCacheTest): Insane
FieldCache usage(s) ***
VALUEMISMATCH: Multiple distinct value objects for
SegmentCoreReader(owner=_0(4.2.1):C12)+course_id
'SegmentCoreReader(owner=_0(4.2.1):C12)'=>'course_id',class
org.apache.lucene.index.SortedDocValues,0.5=>org.apache.lucene.search.FieldCacheImpl$SortedDocValuesImpl#297645694
(size =~ 320 bytes)
'SegmentCoreReader(owner=_0(4.2.1):C12)'=>'course_id',int,org.apache.lucene.search.FieldCache.NUMERIC_UTILS_INT_PARSER=>org.apache.lucene.search.FieldCacheImpl$IntsFromArray#758496471
(size =~ 80 bytes)
*** END testSearchByQuery(com.test.InsaneFieldCacheTest): Insane FieldCache
usage(s) ***
When disabling the group.facet, the test runs as expected.
I found a related issue where the same field was used for the facet and the
group. I verified, the given query does not use the "course_id" field
twice. The error message suggests that sorting on "course_id" was done and
this is stored in the cache.
I set up a copy-field, one for grouping (course_id_grouping) and one for
all other uses (course_id). I got the message again, now for
course_id_grouping. Disabling all caches did not help.
I put a test case on
https://github.com/lischen3229/solrInsaneFieldCacheErrorTest to replicate
the issue.
Any pointers on how to get the facets displaying the group counts instead
of the document counts highly appreciated.
Best,
Elisabeth
