| From: | Peter Geoghegan <pg(at)bowt(dot)ie> |
|---|---|
| To: | Andres Freund <andres(at)anarazel(dot)de> |
| Cc: | Ashwin Agrawal <aagrawal(at)pivotal(dot)io>, Rafia Sabih <rafia(dot)pghackers(at)gmail(dot)com>, PostgreSQL mailing lists <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Zedstore - compressed in-core columnar storage |
| Date: | 2019-04-15 20:07:38 |
| Message-ID: | CAH2-WzmZtqLpqV5hf+GUi3RAreeXC7DPMiFX25VOSakp95v+Qw@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Mon, Apr 15, 2019 at 1:02 PM Andres Freund <andres(at)anarazel(dot)de> wrote:
> There's not much of an alternative currently. Indexes require tid
> looking things, and as a consequence (and some other comparatively small
> changes that'd be required) tableam does too.
I'm trying to establish whether or not that's the only reason. It
might be okay to use the same item pointer struct as the
representation of a integer-like logical identifier. Even if it isn't,
I'm still interested in just how logical the TIDs are, because it's an
important part of the overall design.
> That's one of the reasons why I've been trying to get you to get on
> board with allowing different leaf-level "item pointer equivalents"
> widths inside nbtree...
Getting me to agree that that would be nice and getting me to do the
work are two very different things. ;-)
--
Peter Geoghegan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tomas Vondra | 2019-04-15 20:17:09 | Re: Zedstore - compressed in-core columnar storage |
| Previous Message | Andres Freund | 2019-04-15 20:02:47 | Re: Zedstore - compressed in-core columnar storage |