Re: Re: BUG #12885: The result of casting a double to an integer depends on the database version

>>>>> "Dean" == Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com> writes:

Dean> I'm not convinced that it is a good idea to change the default
Dean> rounding mode for numeric.

Dean> For one thing, numeric is commonly used for monetary data, and
Dean> changing the default rounding mode might well break existing
Dean> applications where rounding is important, and which rely on the
Dean> current behaviour.

On the other hand, we've had complaints (on IRC) from people doing
calculations with monetary data who complain that we don't do
round-to-even (which is widely known as "Banker's Rounding") as their
requirements demand.

To quote the last person who brought it up there:

>> We have strict accounting practice to follow -- and our CFO has
>> instructed to use even-rounding (aka bankers rounding)... and to not
>> use our typical rounding behavior we learned in 1st grade. :)

Dean> Adding a version of round() with support for a choice of various
Dean> rounding modes might be a better way to go.

There's still the issue of casting to ::numeric(m,n) to consider, and
whatever internal rounding is being done in division and numeric
functions.

Dean> Also, if the default is being changed, I think that merits wider
Dean> discussion (on a thread with a more appropriate title) to see if
Dean> it is likely to be an issue for anyone.

Absolutely.

--
Andrew (irc:RhodiumToad)