I love how easy it is to search multiple indices at once in ES, but I
started to wonder: how does scoring work with multiple indices? From what I
understand, the idf of a term within an index was only valid for that
particular index. When you search across multiple indices does that someone
combine the idfs?
Or does the searcher end up just comparing score values (even though they
are strictly comparable), relying on us to provide index boosts to
compensate?
A query to n indices, results in n result sets. Scores can differentiate
within a result set, but not across result sets.
On Friday, November 16, 2012, Andrew O'Brien wrote:
Hi everyone,
I love how easy it is to search multiple indices at once in ES, but I
started to wonder: how does scoring work with multiple indices? From what I
understand, the idf of a term within an index was only valid for that
particular index. When you search across multiple indices does that someone
combine the idfs?
Or does the searcher end up just comparing score values (even though they
are strictly comparable), relying on us to provide index boosts to
compensate?
So when I GET /_search without specifying an index, I get results from multiple indices. I don't have them in front of me now, but I'm pretty sure they were mixed together (which follows from what I understood from the Multi-Tenancy blurb on the ES front page). How does that work?
You can choose how elasticsearch should handle distributed scoring by
specifying search_type for your query:
On Friday, November 16, 2012 9:19:17 AM UTC-5, Andrew O'Brien wrote:
So when I GET /_search without specifying an index, I get results from
multiple indices. I don't have them in front of me now, but I'm pretty sure
they were mixed together (which follows from what I understood from the
Multi-Tenancy blurb on the ES front page). How does that work?
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.