Thread: lc_numeric and negative-prefix
Hi.
We're having this thread over at https://github.com/impossibl/pgjdbc-ng/issues/420
Can anybody shed som light on when negative-prefix is supposed to be respected by PG's formatting-functions?
In lc_numeric='nb_NO.UTF-8' negative-prefix is '−'(8722), not '-'(45), at least in the JAVA-world, but it seems the JDBC-driver must use lc_numeric='C' when parsing server output.
-- 
Andreas Joseph Krogh
CTO / Partner - Visena AS
Mobile: +47 909 56 963
Attachment
På torsdag 26. september 2019 kl. 00:53:28, skrev Tom Lane <tgl@sss.pgh.pa.us>:
Andreas Joseph Krogh <andreas@visena.com> writes:
> Can anybody shed som light
> on when negative-prefix is supposed to be respected by PG's
> formatting-functions? In lc_numeric='nb_NO.UTF-8' negative-prefix is '−'(8722),
> not '-'(45), at least in the JAVA-world, but it seems the JDBC-driver must use
> lc_numeric='C' when parsing server output. -- Andreas Joseph Krogh CTO / Partner
PG does not consider LC_NUMERIC at all when producing output from
the standard numeric data types (and we aren't going to start).
AFAIR the only functions that do pay attention to LC_NUMERIC are
to_char() and friends.
regards, tom lane
Thanks for clarifying.
-- 
Andreas Joseph Krogh
CTO / Partner - Visena AS
Mobile: +47 909 56 963