Обсуждение: pg_dump index/constraint creation order

While watching a 9-hour 60GB network load from 7.4.6 pg_dump into
8.1.2, I noticed the order in which indices and constraints are
created appears to be their creation order.

Would it make more sense to have pg_dump dump indexes grouped by
the table?  That way, if a table got loaded into cache for one
index creation, it might still be there for the immediatly
following index creations on the same table...?

Ed

On May 14, 2006, at 12:27 AM, Ed L. wrote:

> While watching a 9-hour 60GB network load from 7.4.6 pg_dump into
> 8.1.2, I noticed the order in which indices and constraints are
> created appears to be their creation order.

If you use the 8.1.2 pg_dump to make the dump from your 7.4.6 DB,
what is the order like?
I suspect it would be more to what you're expecting.  In general, you
should use the pg_dump  corresponding to the version into which you
are loading for best results.

On Monday May 15 2006 11:14 am, Vivek Khera wrote:
> On May 14, 2006, at 12:27 AM, Ed L. wrote:
> > While watching a 9-hour 60GB network load from 7.4.6 pg_dump
> > into 8.1.2, I noticed the order in which indices and
> > constraints are created appears to be their creation order.
>
> If you use the 8.1.2 pg_dump to make the dump from your 7.4.6
> DB, what is the order like?
> I suspect it would be more to what you're expecting.  In
> general, you should use the pg_dump  corresponding to the
> version into which you are loading for best results.

8.1.2 has same problem (try attached script to see).

Ed