it would be nice to know if there is any performance difference if using a
separate index instead of a simple routing mechanism.
I've got following situation: Two different document types, but same
mapping definition.
So I'm curios which option fits the best for my needs:
Separate index for each document type
Same index, but different index types (with same mappings)
Same Index and same index type, but routing (?routing={DocumentType})
Is there any difference in performance or are there any other advantages or
disadvantages? What would you recommend?
It does not sounds like option 3 (routing) would be the correct one, since
it will force all the documents of a specific type to be in the same shard.
When they exist in the same index, then a query executes against all the
content and filtered (but highly cached) based on the type. If its in a
separate index, then there is no need to filter. So, different indices will
be faster.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.