Postgres Pro
Downloads
Home   >   mailing lists

Re: Index Skip Scan - Mailing list pgsql-hackers

From Andy Fan
Subject Re: Index Skip Scan
Date
Msg-id CAKU4AWqqsagFb=J4bu5YfxokZvmqb_hGDZFtimfSzNLBFthUuQ@mail.gmail.com
Whole thread Raw
In response to Re: Index Skip Scan  (David Rowley <dgrowleyml@gmail.com>)
Responses Re: Index Skip Scan  (David Rowley <dgrowleyml@gmail.com>)
List pgsql-hackers
Tree view

I think the UniqueKeys may need to be changed from using
EquivalenceClasses to use Exprs instead.

When I try to understand why UniqueKeys needs EquivalenceClasses, 
see your comments here.  I feel that  FuncExpr can't be
used to as a UniquePath even we can create unique index on f(a)
and f->strict == true.  The reason is even we know a is not null,
 f->strict = true.  it is still be possible that f(a) == null.  unique index
allows more than 1 null values.  so shall we move further to use varattrno
instead of Expr?  if so,  we can also use a list of Bitmapset to present multi
unique path of a single RelOptInfo. 

pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: [Patch] pg_rewind: options to use restore_command fromrecovery.conf or command line
Next
From: Craig Ringer
Date:
Subject: Re: Proposal: PqSendBuffer removal