Fork me on GitHub
#xtdb
<
2020-02-21
>
sjharms19:02:06

I am interested in storing documents and then doing a full-text search on a field. What advice do people have if Crux is the primary data store?

refset20:02:40

Hi - I believe most people needing full-text search just replicate documents into a separate Elastic Search cluster and use that. There are hypothetical possibilities for embedding single-node lucene indexing into the Crux query engine (for use in Datalog) which would probably work okay for small data sets (<1GB)

refset20:02:47

Are you mainly hoping to consolidate the number of systems to manage, or to combine full-text search with Datalog queries? Or something else?

sjharms20:02:11

Basically Juxt seems like a perfect model for storing a document which may change over time, and to see what changes happened when. One of the fields common in those documents is :body, and for instance I would like to run queries which count keywords or phrases within those bodies

👍 4
Jack Park21:02:34

There is a plugin for Postgres which seamlessly blends elasticsearch such that sql can now be used for full text search; I wonder if something similar can be performed in Crux which extends Datalog queries to deal with full text.

💯 8
4