On Fri, Apr 8, 2016 at 7:23 AM, Daniel Verite <daniel@manitou-mail.org> wrote:
> Alvaro Herrera wrote:
>
>> I wonder if the business of appending values of multiple columns
>> separated with spaces is doing us any good. Why not require that
>> there's a single column in the cell? If the user wants to put things
>> together, they can use format() or just || the fields together. What
>> benefit is there to the ' '? When I ran my first test queries over
>> pg_class I was surprised about this behavior:
>> alvherre=# select * from pg_class
>> alvherre=# \crosstabview relnatts relkind
>
> ISTM that this could be avoided by erroring out for lack of an
> explicit 3rd column as argument. IOW, we wouldn't assume
> that "no column specified" means "show all columns".
>
> About simply ripping out the possibility of having multiple
> columns into cells, it's more radical but if that part turns out to
> be more confusing than useful, I don't have a problem
> with removing it.
>
> The other case of stringing multiple contents into the same cell
> is when different tuples carry (row,column) duplicates.
> I'm not inclined to disallow that case, I think it would go too far
> in guessing what the user expects.
> My expectation for a viewer is that it displays the results as far as
> possible, whatever they are.
> Also, showing such contents in vertically-growing cells as it
> does now allows the user to spot these easily in the grid when
> they happen to be outliers. I'm seeing it as useful in that case.
This seems like it might be converging on some sort of consensus, but
I'm wondering if we shouldn't push it to 9.7, instead of rushing
decisions that we will later have trouble changing on
backward-compatibility grounds.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company