Robert Haas <robertmhaas@gmail.com> writes:
> On Fri, Feb 17, 2017 at 6:51 AM, Greg Stark <stark@mit.edu> wrote:
>> Moreover, it wouldn't be hard to make sum(float4) use a float8 as an
>> accumulator and then cast to float4 for the final state. That would be
>> 100% compatible with the existing behaviour aside from producing more
>> accurate results.
> Sure, but if somebody wants that, they can get it already just by a
> minor change to the SQL. The point is that adding up float4 as float4
> is a reasonable thing to do. Adding it as float8 might also be a
> reasonable thing to do, but nobody's keeping anybody from doing that.
Also, if we changed sum(float4) to work that way, it would become very
hard to sum float4 in float4 --- you'd pretty much have to build your
own aggregate function, which is a lot harder than just inserting a
cast. The argument for changing this boils down to "nobody would ever
want that", which I do not think I buy. It's been a long time since
I studied numerical analysis, but I think there are applications where
you do want to do that.
regards, tom lane
