CREATE STATISTICS documentation bug

Robert Haas <robertmhaas@gmail.com> writes:

P.S. I think the fact that we print "psql: " before the ERROR here is
useless clutter. We didn't do that in v11 and prior and I think we
should kill it with fire.

Agreed, particularly seeing that the error is not originating with
psql; it's just passing it on.

regards, tom lane

On 2019-Jun-14, Tom Lane wrote:

Robert Haas <robertmhaas@gmail.com> writes:

P.S. I think the fact that we print "psql: " before the ERROR here is
useless clutter. We didn't do that in v11 and prior and I think we
should kill it with fire.

Agreed, particularly seeing that the error is not originating with
psql; it's just passing it on.

+1

Proposal: each program declares at startup whether it wants the program
name prefix or not.

--
�lvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

On 2019-Jun-14, Tom Lane wrote:

Robert Haas <robertmhaas@gmail.com> writes:

P.S. I think the fact that we print "psql: " before the ERROR here is
useless clutter. We didn't do that in v11 and prior and I think we
should kill it with fire.

Agreed, particularly seeing that the error is not originating with
psql; it's just passing it on.

+1

Proposal: each program declares at startup whether it wants the program
name prefix or not.

Well, to clarify: I think it's reasonable to include "psql: " if the
message is originating in psql. So I don't think your idea quite
does what we want.

regards, tom lane

On 2019-Jun-14, Tom Lane wrote:

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

Proposal: each program declares at startup whether it wants the program
name prefix or not.

Well, to clarify: I think it's reasonable to include "psql: " if the
message is originating in psql. So I don't think your idea quite
does what we want.

Hmm, it doesn't.

Maybe the error reporting API needs a bit of a refinement to suppress
the prefix for specific error callsites, then?

--
�lvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On Fri, Jun 14, 2019 at 03:23:29PM -0400, Robert Haas wrote:

https://www.postgresql.org/docs/12/sql-createstatistics.html contains
this example command:

CREATE STATISTICS s2 (mcv) ON (a, b) FROM t2;

But that produces:

psql: ERROR: only simple column references are allowed in CREATE STATISTICS

I think the parentheses around (a, b) just need to be removed.

P.S. I think the fact that we print "psql: " before the ERROR here is
useless clutter. We didn't do that in v11 and prior and I think we
should kill it with fire.

I've pushed a fix for the docs issue.

--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On Sat, Jun 15, 2019 at 7:22 PM Tomas Vondra
<tomas.vondra@2ndquadrant.com> wrote:

I've pushed a fix for the docs issue.

Thanks.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On 2019-06-14 23:04, Alvaro Herrera wrote:

On 2019-Jun-14, Tom Lane wrote:

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

Proposal: each program declares at startup whether it wants the program
name prefix or not.

Well, to clarify: I think it's reasonable to include "psql: " if the
message is originating in psql. So I don't think your idea quite
does what we want.

Hmm, it doesn't.

Maybe the error reporting API needs a bit of a refinement to suppress
the prefix for specific error callsites, then?

This was an oversight and has been fixed. (It was masked if you had a
.psqlrc, which is why I never saw it.)

--
Peter Eisentraut http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services