Hi,

I have a table with about 120 million rows. I'd like to index two
columns. Column id1 has about 15,000 distinct values; column id2 has
about 600,000. Most of the queries are:

select <something> from <table> where id1 = ?
select <something> from <table> where id2 = ?
select <something> from <table> where id1 = ? and id2 = ?

I'm thinking for id1, the column with 15,000 values, a bitmap index
would be appropriate given its low relative cardinality. And for id2
and (id1,id2) a B-tree index.

Just looking for a sanity check...thanks.