Re: [gnome-db] Re: Lo que hablamos de los Oid en libgda.



Hi everybody at GNOME-DB list! Nice to meet you!

Well, first of all, unfortunately I'm not an expert neither in Postgresql 
nor in GDA. Nonetheless I'd like to share my experience in trying to deal 
with large files and databases just in case it is useful...

On Thursday, January 02, 2003 at 20:22:55, Gonzalo Paniagua Javier wrote:
> Yes, this is tightly coupled to postgresql large objects interface and
> that's why it's fine for postgresql...
> 
> ...but, for example, BLOBs in MySQL are closer to postgresql 'bytea'
> (binary strings) type than to postgresql blobs and it wouldn't make

Yes, first I didn't like postgresql's large objects interface and preferred
plain bytea, but inserting big files into a bytea was discouraging: a 3
Mbytes file, with non-isprint(3) bytes properly escaped and inserted into a
Postgresql DB using GDA made a 48 Mbytes process!, and Postgresql was not
very happy either: it took 98% of the CPU for 4-5 minutes to process the 
query and insert the file into the database... :-(

That's why I thought the way to go was using OIDs.

But I don't know if your current work has solved this kind of performance
problems. If it has, great!, I'm eager to use it!

Thank you for all, and if there's something I can help with, testing and
the like, please, tell me and I'll happily do it.

Cheers,

Juan-Mariano.



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]