Hi, I am new to Hibernate Search and was wondering if below usecase is possible.

We have a table that stores transactional messages (xml). Since message sizes can be huge at times, we split messages into multiple records, each record having the same identifier but different sequence number.

So given this usecase, is it possible for Hibernate Search to aggregate these records together while performing the search as they share the same identifier? This will be required due to our chunking approach, as a word of importance can be split into multiple records (For example: "hello world" can be chunked as "hello w" and "orld" records. And we need to match these records if search text is "world").

Thanks.

Hi,
it depends on your mapping. Hibernate Search expects unique entities to be indexed, but it would be fine for example if you have a "Record 1" entity having a relation to all it's "chunk 1 for record 1" and "chunk 2 for record 1".

In such a case, just index the "Record 1" which represents the aggregation of your message (and has a reference to all records chunks), and have it index all chunks as a single collection.

Thanks Sanne, will check this one out.