Re: Searching for a user across two tables taking 23 seconds

[Marek Podmaka]

I think the main problem is:
status like '%active%'
This can't use any indexes. What are the possible values of field status? If it is either 'active' or 'inactive', then you should use status = 'active' and it would use index on this field. But even better is to use status as tinyint (or char(1)) and have values of 0/1.

Try to read manual about the EXPLAIN, it is very useful. Look at the column with rows produced. For your query, it should read 3000 rows from one of the tables and 1 row from another table (= for each row from 1st). 3000*1=3000 rows total which get produced and needs to be analyzed.