Horton

Re: Binary support for pgoutput plugin

Chapman Flack <chap@anastigmatix.net>

View thread View on postgr.es
From: Chapman Flack <chap@anastigmatix.net>
To: Dave Cramer <davecramer@gmail.com>, Andres Freund <andres.freund@enterprisedb.com>
Cc: David Fetter <david@fetter.org>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2019-06-04T21:33:28Z
Lists: pgsql-hackers 
On 6/4/19 4:39 PM, Dave Cramer wrote:
> I haven't really thought this through completely but one place JDBC has
> problems with binary is with
> timestamps with timezone as we don't know which timezone to use. Is it safe
> to assume everything is in UTC
> since the server stores in UTC ?

PL/Java, when converting to the Java 8 java.time types (because those
are sane), will turn a timestamp with timezone into an OffsetDateTime
with explicit offset zero (UTC), no matter what timezone may have been
used when the value was input (as you've observed, there's no way to
recover that). In the return direction, if given an OffsetDateTime
with any nonzero offset, it will adjust the value to UTC for postgres.

So, yes, say I.

Regards,
-Chap

Commits

  1. Weaken type-OID-matching checks in array_recv and record_recv.

  2. Fix construction of updated-columns bitmap in logical replication.

  3. Allow logical replication to transfer data in binary format.

  4. Phase 2 pgindent run for v12.