I have a db of about 250G. In it is a table of Internet domain information containing, among other things, the domain name. In that table I have a serial int index as the primary key to use as a foreign key in a number of other tables. The domain name is a text field. There are about 60M rows in the table, all with unique names.
Many queries begin with the specification of the domain name and then fan out through joins to other tables. Obviously I need to be able to locate a domain name quickly so I can get the index of the row to use in the joins. I believe (but don't know) that a hash index would be better for this than a btree. I don't know enough to have an opinion wrt GiST or GIN indexes. Any advice on this problem would be greatly appreciated.
