Shreya Bhargava <shreya_bhargav@yahoo.com> writes:
> for each of the above value, I created a hash index on a 1.17GB table.
Hash index build doesn't have any need for workspace, so it will not go
faster with larger work_mem. Try it with a regular btree index.
> Am i missing something?
Have you read the statements in the manual that discourage you from
using hash indexes at all? Our current implementation of them has
no known advantages over a btree, and several major disadvantages.
regards, tom lane
