On Jun 22, 2006, at 9:39 PM, Tom Lane wrote:
> Jim Nasby <jnasby@pervasive.com> writes:
>> On Jun 21, 2006, at 12:00 PM, Tom Lane wrote:
>>> This could be avoided by using COPY BINARY format, but I don't
>>> see any
>>> very nice way to do that in the context of pg_dump --- it needs to
>>> intermix COPY data with SQL commands ...
>
>> Would the tar or custom format allow for this? IIRC, at least tar
>> puts all the copied data into separate files...
>
> Well, you could sorta do that, but the case that would stop working is
> pg_restore output to a plain text SQL script (and related issues
> such as
> the ability to use the feature in the context of pg_dumpall). Having
> just gotten done fixing similar inconsistencies in pg_dump/pg_restore
> for BLOBs, I'm not eager to re-introduce 'em for COPY BINARY ...
Yeah, but how many people actually do that anyway? I can't really
come up with a use-case for it, and I'm pretty sure there's other
gains to be had by turning custom or tar format into more of a
'binary dump'. For one thing, that should ease the need to run the
newer version of pg_dump when upgrading (if we put the requisite
brains into pg_restore).
I suppose we could put support in pg_restore to convert between
BINARY and escaped as needed; or just disallow pg_restore from
dumping SQL if there's binary data (maybe have it include copy
statements that reference the specific files).
--
Jim C. Nasby, Sr. Engineering Consultant jnasby@pervasive.com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461