This query requires that you index
children and parent docs as a single block, using the
IndexWriter#updateDocuments API. In each block, the
child documents must appear first, ending with the parent
document. At search time you provide a Filter
identifying the parents, however this Filter must provide
BitSet per sub-reader.
Once the block index is built, use this query to wrap
any sub-query matching only child docs and join matches in that
child document space up to the parent document space.
You can then use this Query as a clause with
other queries in the parent document space.
ToChildBlockJoinQuery if you need to join
in the reverse order.
The child documents must be orthogonal to the parent
documents: the wrapped child query must never
return a parent document.
If you'd like to retrieve
TopGroups for the
resulting query, use the
Note that this is not necessary, ie, if you simply want
to collect the parent documents and don't need to see
which child documents matched under that parent, then
you can use any collector.
NOTE: If the overall query contains parent-only
matches, for example you OR a parent-only query with a
joined child-only query, then the resulting collected documents
will be correct, however the
TopGroups you get
ToParentBlockJoinCollector will not contain every
child for parents that had matched.
org.apache.lucene.search.join for an