Re: [patch] Proposal for \crosstabview in psql

Pavel Stehule wrote:

> > FWIW I think the general idea of this feature (client-side resultset
> > "pivoting") is a good one, but I don't really have an opinion regarding
> > your specific proposal.  I think you should first seek some more
> > consensus about the proposed design; in your original thread [1] several
> > guys defended the idea of having this be a psql feature, and the idea of
> > this being a parallel to \x seems a very sensible one,

Sorry, I meant \q here, not \x.

> > but there's really been no discussion on whether your proposed "+/-"
> > syntax to change sort order makes sense, for one thing.
> 
> I am sorry, but I disagree - the discussion about implementation was more
> than two months, and I believe so anybody who would to discuss had enough
> time to discuss. This feature and design was changed significantly and
> there was not anybody who sent feature design objection.

I just rechecked the thread.  In my reading, lots of people argued
whether it should be called \rotate or \pivot or \crosstab; it seems the
\crosstabview proposal was determined to be best.  I can support that
decision.  But once the details were discussed, it was only you and
Daniel left in the thread; nobody else participated.  While I understand
that you may think that "silence is consent", what I am afraid of is
that some committer will look at this two months from now and say "I
hate this Hcol+ stuff, -1 from me" and send the patch back for syntax
rework.  IMO it's better to have more people chime in here so that the
patch that we discuss during the next commitfest is really the best one
we can think of.

Also, what about the business of putting "x" if there's no third column?
Three months from now some Czech psql hacker will say "we should use
Unicode chars for this" and we will be forever stuck with \pset
unicode_crosstab_marker to change the character to a ☑ BALLOT BOX WITH
CZECH.  Maybe we should think that a bit harder -- for example, what
about just rejecting the case with no third column and forcing the user
to add a third column with the character they choose?  That way you
avoid that mess.

> This feature has only small relation to SQL PIVOTING feature - it is just
> form of view and I am agree with Daniel about sense of this feature.

Yes, I don't disagree there.  Robert Haas and David Fetter also
expressed their support for psql-side processing, so I think we're good
there.

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