Re: ICU for global collation

On 04.01.22 17:03, Peter Eisentraut wrote:
>> There are really a lot of places with this new code.  Maybe it could
>> be some
>> new function/macro to wrap that for the normal case (e.g. not
>> formatting.c)?
>
> Right, we could just put this into pg_newlocale_from_collation(), but
> the comment there says
>
>  * In fact, they shouldn't call this function at all when they are dealing
>  * with the default locale.  That can save quite a bit in hotspots.
>
> I don't know how to assess that.

I tested this a bit. I used the following setup:

create table t1 (a text);
insert into t1 select md5(generate_series(1, 10000000)::text);
select count(*) from t1 where a > '';

And then I changed in varstr_cmp():

if (collid != DEFAULT_COLLATION_OID)
mylocale = pg_newlocale_from_collation(collid);

to just

mylocale = pg_newlocale_from_collation(collid);

I find that the \timing results are indistinguishable. (I used locale
"en_US.UTF-8" and made sure that that code path is actually hit.)

Does anyone have other insights?