On 10 Sep 2003 at 22:44, Tom Lane wrote: > James Robinson <[EMAIL PROTECTED]> writes: > > Is this just a dead end, or is there some variation of this that might > > possibly work, so that ultimately an undoctored literal number, when > > applied to an int8 column, could find an index? > > I think it's a dead end. What I was playing with this afternoon was > removing the int8-and-int4 comparison operators from pg_operator. > It works as far as making "int8col = 42" do the right thing, but I'm > not sure yet about side-effects.
Is it possible to follow data type upgrade model in planner? Something like in C/C++ where data types are promoted upwards to find out better plan? int2->int4->int8->float4->float8 types. That could be a clean solution.. just a thought.. Bye Shridhar -- Hlade's Law: If you have a difficult task, give it to a lazy person -- they will find an easier way to do it. ---------------------------(end of broadcast)--------------------------- TIP 8: explain analyze is your friend