Optimizing a large mySQL table

Question

I have a table of location codes that is about 2.3 million rows, with no indices. My site needs to routinely query it for things like distinct state codes, counties filtered down by state, cities filtered down by county and/or state, etc... The problem is that because it is so big and that it has now indices, queries can be slow.

My question is, should I make indices on the columns and various combinations of columns I'd be querying on? What indices should I make? I'm imagining these indices:

• Unique/primary index on feature_id (obviously)
• Index on state_numeric
• Index on state_numeric, county_numeric

Here's a link to the image of the table

Answer 1

It would be best to make individual indexes on each of the commonly used fields. Note that updates and inserts will be a bit slower, so don't make more indexes than needed. If you are running a lot of queries that routinely use specific combinations of fields, then composite indexes on those would be good as well.