Re: [HACKERS] Upgrades for 6.4.1

"Thomas G. Lockhart" <lockhart(at)alumni(dot)caltech(dot)edu> writes:
>> * SELECT DISTINCT i FROM dtest ORDER BY j generates strange output

> In my simple test case, it orders by j, then only shows i. Is that
> strange?

The thing that is "strange" is that you get nonunique values of i,
which is definitely a bit unexpected for "SELECT DISTINCT":

play=> SELECT * FROM dtest;
i| j
-+--
1|10
1|11
1|12
2|20
3|30
(5 rows)

play=> SELECT DISTINCT i FROM dtest ORDER BY j;
i
-
1
1
1
2
3
(5 rows)

The reason that this is happening is that the "distinct" filter is
actually being run on i,j not just i (see "distinct + order by" thread
in the hackers archives around 8-Nov-98).

I don't know whether the SQL standard defines how this combination of
features ought to work ... but our current behavior seems fairly
surprising...

>> * Allow constraint NULL just as we honor NOT NULL

> Fundamental yacc problem with this as I recall. Gives rise to
> shift/reduce problems since it is ambiguous with other uses of "NULL" in
> the same area.

More to the point, what possible use would a column constrained to NULL
be? Might as well just not have it in the table...

regards, tom lane