Answer the question
In order to leave comments, you need to log in
How to do a quick full text search?
For a project, some analogue of stackoverflow.com, only on mobile topics - http://www.uagadget.com , it is required to implement a full-text search for questions and answers.
Server capacities are currently not great, as is often the case with all startups. Therefore, I want to ask you, what is the best full-text search engine to use? It is necessary that the search be fast and not very demanding on resources (in particular, on RAM). Search by native PostgreSQL tools, I think, can be quite "cargo".
What are your opinions?
Answer the question
In order to leave comments, you need to log in
Consider Sphinx as an option . Fast index building, fast search.
More: habrahabr.ru/hub/sphinx/posts/
And also a comparison of full-text search engines .
Generally speaking, PostgreSQL's built-in full-text search is quite fast. That is, brakes under the boundary conditions of searching for 1,000,000 documents (the volume of each is from 1,000 to 300,000 characters, the average document size is 2,000 characters) are generally not observed. In any case, on a dedicated server (in this particular case, Hetzner is used with 24Gb of memory, of which PostgreSQL eats about a quarter, the PostgreSQL version is 9.2 - the results were slightly different on earlier versions.)
I would recommend trying to test your current configuration for performance with different engines. And, in the case of using postgres, do not use highlighting results.
Look towards Solr. Regarding system requirements and calculation of resources, you can read here and here + calculation in excel
I have a PostgreSQL + Sphinx bundle on my project. Everything works great, it does not eat a lot of resources.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question