Re: Invalid field size

On 07/04/2017 08:37 AM, Moreno Andreo wrote:
> Il 04/07/2017 17:25, Tom Lane ha scritto:
>> Moreno Andreo <moreno(dot)andreo(at)evolu-s(dot)it> writes:
>>> Il 04/07/2017 16:51, Tom Lane ha scritto:
>>>> Pushing binary data around on Windows is always a hazardous
>>>> proposition.
>>> So what you are saying is "in the last 5 years you've been extremely
>>> lucky?" :-)
>> Yup, particularly now that you mention moving the files between machines.
>> What did you do that with exactly?
> Trying to answer your question (I hope I understood correctly, English
> is not my mother tongue)
> What I do is, given a database, to COPY every table to a file, and then
> pack them up in one with a zip (except this table, that's been excluded
> from compression for its size and consequent compression time), so my
> backup is made up by 2 files, one with "normal data" and one with the
> result of COPYing this table to a file.
>
> A question that comes while I'm writing: but pg_dump with custom format
> is not using COPY with binary format?

A quick look through the source indicates to me that it is not using
BINARY. Then again I am not a C programmer, so take that into account.
It would stand to reason that it would not use BINARY as using
pg_dump/pg_restore is supposed to be portable across OS, machine
architecture and to a certain degree Postgres versions. COPY WITH BINARY
would work against that:

https://www.postgresql.org/docs/9.1/static/sql-copy.html

"The binary format option causes all data to be stored/read as binary
format rather than as text. It is somewhat faster than the text and CSV
formats, but a binary-format file is less portable across machine
architectures and PostgreSQL versions.

>
> Thanks
> Moreno
>
>
>

--
Adrian Klaver
adrian(dot)klaver(at)aklaver(dot)com