When I first tried mlocate, I was amazed at two things: First: How quick it was (almost instantaneous). And then: how much data it gave back (way too much to be useful).

I suppose that this is old hat to some of you old timers, but I finaly learned to use it with grep. Major improvement in narrowing the search. For Instance: "mlocate -i index | grep htm" will really cut down on the number of lines returned.

Another problem that bothered me was getting the gui file manager to interface with the information from mlocate. In most of them you can't just cut and paste the name of a long directory from the Terminal into the gui file manager. But I found a way around that, too: Just highlight the pathname of the directory that you want displayed in your file manager, and while still in your terminal, type the name of your file browser followed by a space, and then tap the middle mouse button and enter. For example, using file manager 'caja': "caja <tapa> /path/name/of/directory".

Or you can enter a document in Libreoffice in a similar way. Just highlight the comoplete pathname of the document file in the list that you have on your terminal from your mlocate search, and then go to Libreoffice writer:file>open and click on the small icon of a pencil. in the line that appears, tap the middle mouse button.

Once you get stated you'll find lots of other advavatages.

Hopefully you have read the manpage for [m]locate - it supports globbing in addition to both basic and extended regex. You can control to a significant degree how much output you get. But there are usualy still cases which require filtering as you suggested.
The file list used is a database created by updatedb, hence the speed advantage over tools like "find"