I've been reading the postgresql manual and I find there is very little
discussion about hash compared to btree. Most of the focus seems to be on
using btree indices even that the default for 'create index' is btree
also. From the documentation, it seems the only difference between either
searching method is that btree can be used with multiple operators whilst
hash can only be used with '='. Furthermore, hash seems to be contained in
memory, so should be limited to small queries or, in my case, queries
using limit (without using sort which would need to retrieve the entire
data anyways).
My conclusion is that if I can live with just using '=' and using slightly
more memory, I should be using hash. Unfortunately, there is very little
sign in the documentation that I should be using hash at all. Perhaps I
have missed something important.
If someone could help me make a more rational decision on using searching
methods, I'd appreciate.
Thanks in advance,
Marc