Yeah - it can make it easier to implement transactional semantics by
storing them in the database, although for simple operations it wouldn't
be hard to replicate this manually. And you are going to incur a
performance penalty by storing them in the database.
Another thing to consider is that storing them in the file system makes
it much easier to browse the images using third-party tools, update
them, archive them (by gzipping or whatever). This is much more
difficult if they are stored in the database.
On Fri, 5 Jan 2007 15:51:59 -0600, "Bruno Wolff III" <bruno@wolff.to>
said:
> On Fri, Jan 05, 2007 at 15:26:45 -0500,
> James Neff <jneff@tethyshealth.com> wrote:
> > "... and Moses said unto them, 'The eleventh commandment : thou shalt
> > store images in a database!'..."
> >
> > What if you had another database where you stored just the images and
> > not back it up if you don't want to?
>
> I think the main reason to keep images in the database is if you need
> transactional semantics. If you are updating images and transactions that
> started before the update, need to see the old version you are going to
> want them in the database. I suspect this need isn't very common though.
>
> ---------------------------(end of broadcast)---------------------------
> TIP 3: Have you checked our extensive FAQ?
>
> http://www.postgresql.org/docs/faq