Re: Marking indexes out of date (WAS: loading data, - Mailing list pgsql-general

From Bruce Momjian
Subject Re: Marking indexes out of date (WAS: loading data,
Date
Msg-id 200612132306.kBDN67F11744@momjian.us
Whole thread Raw
In response to Re: Marking indexes out of date (WAS: loading data, creating  (Glen Parker <glenebob@nwlink.com>)
List pgsql-general
Glen Parker wrote:
> Martijn van Oosterhout wrote:
> > On Fri, Dec 08, 2006 at 12:49:30PM -0800, Glen Parker wrote:
> >> I'd like to see a general way to take indexes off line without actually
> >> losing their definitions.  For example, something like "ALTER TABLE [EN
> >> | DIS] ABLE INDEXES", "ALTER INDEX [EN | DIS] ABLE", etc.  This could
> >> also be used internally when a backend encounters an error
> >> reading/writing an index.  Rather than refusing to execute queries, it
> >> could just ignore indexes it knows are disabled or bad in some way and
> >> re-plan as needed.
> >
> > One issue would be that even disabled indexes would need to be updated
> > when there are new rows. If you don't update the index when it's
> > disabled, then re-enabling will essentially need to rebuild the index.
>
>
> That's what I had in mind.  You could just as easily blow away the index
> file(s).  It's just that I don't want it to toss the index *definition*.

Well, you can just pg_dump the index definition to a file.  What more
would someone want?

--
  Bruce Momjian   bruce@momjian.us
  EnterpriseDB    http://www.enterprisedb.com

  + If your life is a hard drive, Christ can be your backup. +

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: creating an index with tablespace on a dedicated filesystem (postgresql 8.1.5)
Next
From: Manuel Sugawara
Date:
Subject: Slow query in 8.2.0