Re: [HACKERS] Backup database with entries > 8192 KBytes
| От | Tom Lane |
|---|---|
| Тема | Re: [HACKERS] Backup database with entries > 8192 KBytes |
| Дата | |
| Msg-id | 1279.933804865@sss.pgh.pa.us обсуждение исходный текст |
| Ответ на | Backup database with entries > 8192 KBytes ("G. Anthony Reina" <reina@nsi.edu>) |
| Список | pgsql-hackers |
"G. Anthony Reina" <reina@nsi.edu> writes:
> I know that I can't insert a tuple into Postgres > 8192 KBytes long.
Er, make that 8K. If the limit were 8Meg, a lot fewer people would be
complaining about it...
> need to store data in a variable length float array which can take up a
> total length of greater than this amount. To get around the limit, we
> simply insert a zeroed array (which takes up less character space)
Huh? I assume you're talking about an array of float4 (or float8).
That's going to be 4 (or 8) bytes per value internally, zero or not.
Maybe you are thinking of the external textual representation of the
array.
> Okay, long intro for a short question. When we do a pg_dump and then
> restore the database should the COPY contained within the pg_dumped file
> be able to handle these long arrays?
Offhand I'm not sure. I don't see any obvious restriction in copy.c,
but there could be lurking problems elsewhere. Have you tried it?
regards, tom lane
В списке pgsql-hackers по дате отправления: