Re: Extended Statistics set/restore/clear functions. - Mailing list pgsql-hackers

From Corey Huinker
Subject Re: Extended Statistics set/restore/clear functions.
Date
Msg-id CADkLM=dZr_Ut3jKw94_BisyyDtNZPRJWeOALXVzcJz=ZFTAhvQ@mail.gmail.com
Whole thread Raw
In response to Re: Extended Statistics set/restore/clear functions.  (Michael Paquier <michael@paquier.xyz>)
Responses Re: Extended Statistics set/restore/clear functions.
List pgsql-hackers
So, I have done more work on all that, and extracted two more pieces,
applying them:
- The bits for dependencies in the restore function, .
- The pg_dump bits for dependencies and n_distinct, as of
c32fb29e979d.  It was sad here that your posted patch did not use the
keys of statistics_format.h.  A comment was also missing in
AdjustUpgrade.pm, where the code will be able to handle the "version"
tag included in the dumps.

Hooray!
 
Another thing that I strongly suspect is going to break in the
buildfarm are the cross-version upgrades, because the format of
the pg_dependencies and pg_ndistinct is not completely portable.

We had a similar rocky experience with pg_restore_relation_stats(), so your concerns are warranted.
 
This is a filter similar to the "version" field, except that we would
replace all the contents of the pg_ndistinct and pg_dependencies
values inside single quotes, protecting the dump comparisons for the
new and old versions.  Let's see first what the buildfarm says and if
I need to pull this switch.

+1 
 
And it means that at this point we are feature-complete for these two
stats kinds, so there will be at least something for this release even
if MCV and expressions don't make it.  Still, the other pieces should
be slightly easier to integrate now that the core parts have been
added.
--
Michael

And here's a rebased set

0001 - adds a test left out of dependencies
0002 - mcv (plus pgdump)
0003 - exprs (plus pgdump)
0004 - capstone set-difference test that could not have succeeded until all stat kinds were implemented.
 
Attachment

pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: Newly created replication slot may be invalidated by checkpoint
Next
From: Zsolt Parragi
Date:
Subject: Re: tablecmds: reject CLUSTER ON for partitioned tables earlier