Обсуждение: Problem with multiple SUMs

Поиск
Список
Период
Сортировка

Problem with multiple SUMs

От
"D'Arcy" "J.M." Cain
Дата:
After recent changes I find an error with SUM when summing more than
one column.  Here is the test sequence.

DROP TABLE x;
CREATE TABLE x (a int, b int); 
INSERT INTO x VALUES (1, 5);
INSERT INTO x VALUES (2, 7);
SELECT * FROM x;
SELECT SUM(a) FROM x;
SELECT SUM(b) FROM x;
SELECT SUM(a), SUM(b) FROM x;

The last three statements give the following expected results when
run on a system compiled Jan 19.

darcy=> SELECT SUM(a) FROM x;
sum
--- 3
(1 row)

darcy=> SELECT SUM(b) FROM x;
sum
---12
(1 row)

darcy=> SELECT SUM(a), SUM(b) FROM x;
sum|sum
---+--- 3| 12
(1 row)

On a system compiled Jan 27, I see the following.

darcy=> SELECT SUM(a) FROM x;
sum
--- 3
(1 row)

darcy=> SELECT SUM(b) FROM x;
sum
---12
(1 row)

darcy=> SELECT SUM(a), SUM(b) FROM x;
sum|sum
---+---12| 12
(1 row)

See how the individual sums are correct but I can no longer get both
sums in one select.

-- 
D'Arcy J.M. Cain <darcy@{druid|vex}.net>   |  Democracy is three wolves
http://www.druid.net/darcy/                |  and a sheep voting on
+1 416 424 2871     (DoD#0082)    (eNTP)   |  what's for dinner.

Re: [HACKERS] Problem with multiple SUMs

От
Bruce Momjian
Дата:
I am working on it.


> After recent changes I find an error with SUM when summing more than
> one column.  Here is the test sequence.
> 
> DROP TABLE x;
> CREATE TABLE x (a int, b int); 
> INSERT INTO x VALUES (1, 5);
> INSERT INTO x VALUES (2, 7);
> SELECT * FROM x;
> SELECT SUM(a) FROM x;
> SELECT SUM(b) FROM x;
> SELECT SUM(a), SUM(b) FROM x;
> 
> The last three statements give the following expected results when
> run on a system compiled Jan 19.
> 
> darcy=> SELECT SUM(a) FROM x;
> sum
> ---
>   3
> (1 row)
> 
> darcy=> SELECT SUM(b) FROM x;
> sum
> ---
>  12
> (1 row)
> 
> darcy=> SELECT SUM(a), SUM(b) FROM x;
> sum|sum
> ---+---
>   3| 12
> (1 row)
> 
> On a system compiled Jan 27, I see the following.
> 
> darcy=> SELECT SUM(a) FROM x;
> sum
> ---
>   3
> (1 row)
> 
> darcy=> SELECT SUM(b) FROM x;
> sum
> ---
>  12
> (1 row)
> 
> darcy=> SELECT SUM(a), SUM(b) FROM x;
> sum|sum
> ---+---
>  12| 12
> (1 row)
> 
> See how the individual sums are correct but I can no longer get both
> sums in one select.
> 
> -- 
> D'Arcy J.M. Cain <darcy@{druid|vex}.net>   |  Democracy is three wolves
> http://www.druid.net/darcy/                |  and a sheep voting on
> +1 416 424 2871     (DoD#0082)    (eNTP)   |  what's for dinner.
> 
> 


--  Bruce Momjian                        |  http://www.op.net/~candle maillist@candle.pha.pa.us            |  (610)
853-3000+  If your life is a hard drive,     |  830 Blythe Avenue +  Christ can be your backup.        |  Drexel Hill,
Pennsylvania19026

Re: [HACKERS] Problem with multiple SUMs

От
Tom Lane
Дата:
"D'Arcy" "J.M." Cain <darcy@druid.net> writes:
> After recent changes I find an error with SUM when summing more than
> one column. ...
> See how the individual sums are correct but I can no longer get both
> sums in one select.

Good eye!

Actually, it looks like *any* two aggregates conflict --- we're
reporting the result of the rightmost aggregate for all aggregate
functions in a SELECT.  Using D'Arcy's test table, I also tried

treetest=> SELECT AVG(a), SUM(a) FROM x;
avg|sum
---+--- 3|  3
(1 row)

treetest=> SELECT AVG(a), SUM(b) FROM x;
avg|sum
---+---12| 12
(1 row)

treetest=> SELECT AVG(a), COUNT(b) FROM x;
avg|count
---+----- 2|    2
(1 row)

Oops.

This bug appears to explain some of the regression-test failures I'm
seeing --- numerology and select_having both contain multiple-aggregate
commands that are failing.

In the select_having test, it looks like multiple aggregates used in
the HAVING clause of a SELECT are suffering the same sort of fate
as those in the target list.
        regards, tom lane

Re: [HACKERS] Problem with multiple SUMs

От
Bruce Momjian
Дата:
Fixed.


> "D'Arcy" "J.M." Cain <darcy@druid.net> writes:
> > After recent changes I find an error with SUM when summing more than
> > one column. ...
> > See how the individual sums are correct but I can no longer get both
> > sums in one select.
> 
> Good eye!
> 
> Actually, it looks like *any* two aggregates conflict --- we're
> reporting the result of the rightmost aggregate for all aggregate
> functions in a SELECT.  Using D'Arcy's test table, I also tried
> 
> treetest=> SELECT AVG(a), SUM(a) FROM x;
> avg|sum
> ---+---
>   3|  3
> (1 row)
> 
> treetest=> SELECT AVG(a), SUM(b) FROM x;
> avg|sum
> ---+---
>  12| 12
> (1 row)
> 
> treetest=> SELECT AVG(a), COUNT(b) FROM x;
> avg|count
> ---+-----
>   2|    2
> (1 row)
> 
> Oops.
> 
> This bug appears to explain some of the regression-test failures I'm
> seeing --- numerology and select_having both contain multiple-aggregate
> commands that are failing.
> 
> In the select_having test, it looks like multiple aggregates used in
> the HAVING clause of a SELECT are suffering the same sort of fate
> as those in the target list.
> 
>             regards, tom lane
> 
> 


--  Bruce Momjian                        |  http://www.op.net/~candle maillist@candle.pha.pa.us            |  (610)
853-3000+  If your life is a hard drive,     |  830 Blythe Avenue +  Christ can be your backup.        |  Drexel Hill,
Pennsylvania19026