Re: Proposal for internal Numeric to Uint64 conversion function.

Amul Sul <sulamul(at)gmail(dot)com> writes:
> On Wed, Feb 16, 2022 at 4:50 PM Peter Eisentraut
> <peter(dot)eisentraut(at)enterprisedb(dot)com> wrote:
>> On 16.02.22 06:00, Amul Sul wrote:
>>> Currently, numeric_pg_lsn is the only one that accepts the Numeric
>>> value and converts to uint64 and that is the reason all the type
>>> conversion code is embedded into it.

>> There are other functions such as numeric_int8() that work similarly.
>> If you are going to refactor, then they should all be treated similarly.
>> I'm not sure if it's going to end up being beneficial.

> Yeah, that's true, I am too not sure if we really need to refactor
> all those; If we want, I can give it a try.

There are several places that call numeric_int8, and right now they
have to go through DirectFunctionCall1, which is ugly and inefficient.
I think a refactoring that exposes some more-convenient API for that
could be useful. The patch as-presented isn't very compelling for
lack of callers of the new function; but if it were handling the
int64 and uint64 cases alike, and maybe the float8 case too, that
would seem more convincing. We already expose APIs like int64_to_numeric,
so the lack of similar APIs for the other direction seems odd.

It also feels to me that numeric_pg_lsn(), per se, doesn't belong in
numeric.c. A pretty direct comparison is numeric_cash(), which is
not in numeric.c but cash.c.

regards, tom lane