Hi All,
Having pg_dump dump tables in clustered index order is something we've
found we've needed a fair number of times (for ex. when copying a large
logging tables or sets of tables out of one database where the order is not
maintained into another for running a bunch of backend analysis) as it
saves us the clustering step which is often longer then the copy step
itself.
I wanted to gauge the interest in adding an option for this to pg_dump. A
(not production ready) patch that we've been using off of the 9.1.2 tag to
implement this is attached or can be viewed
here<https://github.com/tgarnett/postgres/commit/d4412aa4047e7a0822ee93fa47a1c0d282cb7925>.
It adds a --cluster-order option to pg_dump. If people have any
suggestions on better ways of pulling out the order clause or other
improvements that would be great too.
Tim
Responses
pgsql-hackers by date
| Next: | From: Peter Geoghegan | Date: 2012-02-22 23:17:53 |
| Subject: Re: 16-bit page checksums for 9.2 |
| Previous: | From: Robert Haas | Date: 2012-02-22 22:57:05 |
| Subject: Re: WIP: proof of concept patch for fixing quantified regex backrefs |