An index across the complete table will eat up lots of disk space, and also will make each insert slow(er). Think of an index as (somewhat structured) hash keys, where you can also quickly look up strings starting with a specific string.

If your table is largely read-only and you have the disk space to spare, the index will speed up queries maybe a bit more than an index only on the primary keys, because the DB can satisfy the query completely from the index without hitting the table storage.

The triggers will slow down all UPDATE statements but leave SELECT untouched and also will not use additional storage.

Ok, thank you! I do think that since the database will probably grow relatively large, I don't want to be too wasteful with the space, and slightly slower imports are probably fine. I'll try to get the trigger working and maybe I'll do a comparison with the unique key method too.