Elasticsearch over Postgres?

Question

PostgreSQL

My new project has a database that receives data from another service. There was a task to implement a full-fledged search for all attributes with many criteria, so I chose Elasticsearch. It is not possible to change the way the data is received (immediately to the elastic). For indexing tables, I use https://github.com/jprante/elasticsearch-jdbc , but unfortunately nesting is not implemented when accessing multiple tables (the values ​​of old arrays are duplicated https://github.com/jprante/elasticsearch-jdbc/ issu... ). So far, only one solution is on my mind - to index all the necessary tables, and collect the result together in several queries. Maybe someone came across or come up with something else? Thanks in advance!

Answer 1

I think it's best to avoid the extra. tool, if possible. If possible, it is better to use the full-text search of PG. These opportunities are now quite wide. We ourselves use a bunch of sphinx + pg, and because of this we have a lot of problems. But we have a large and complex system, with a serious search, and it has been developing for a long time.