Postgres Pro
Downloads
Home   >   mailing lists

Thread: count different values in column

count different values in column

From
"Albrecht Berger"
Date:
Hello,
today I have a new question :)

I've a resulttable is generated by a couple of subselects and joins which
contains only one column.

Now I want to count the different values in the column.

Resulttable :  a
------
10
12
12
12
14
14
15

I need the value which is repeated most in the resulttable (in this example
12 ) !


Any suggestions ?


Thx
berger

Re: count different values in column

From
Tod McQuillin
Date:
On Fri, 19 Apr 2002, Albrecht Berger wrote:

> Now I want to count the different values in the column.
>
> Resulttable :
>    a
> ------
> 10
> 12
> 12
> 12
> 14
> 14
> 15
>
> I need the value which is repeated most in the resulttable (in this example
> 12 ) !

SELECT a, count(a) FROM resulttable GROUP BY a ORDER BY 2 DESC
-- 
Tod McQuillin

Re: count different values in column

From
"Gautham S. Rao"
Date:
Hi Berger,

Try,
select a from  Resulttable group by a having count(a) = (select max(cnt)
from (select count(a) as cnt from  Resulttable group by a) as tmp);

Regards,
Gautham.

----- Original Message -----
From: "Albrecht Berger" <berger1517@gmx.ch>
To: "pgsql" <pgsql-sql@postgresql.org>
Sent: Friday, April 19, 2002 3:00 PM
Subject: [SQL] count different values in column


> Hello,
> today I have a new question :)
>
> I've a resulttable is generated by a couple of subselects and joins which
> contains only one column.
>
> Now I want to count the different values in the column.
>
> Resulttable :
>    a
> ------
> 10
> 12
> 12
> 12
> 14
> 14
> 15
>
> I need the value which is repeated most in the resulttable (in this
example
> 12 ) !
>
>
> Any suggestions ?
>
>
> Thx
> berger
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster
>

Re: count different values in column

From
Christoph Haller
Date:
select max(count) from (
select a, count(a) from Resulttable group by a 
) as foo; max
-----  3
(1 row)
should do it. 
Regards, Christoph 

> Hello,
> today I have a new question :)
> 
> I've a resulttable is generated by a couple of subselects and joins which
> contains only one column.
> 
> Now I want to count the different values in the column.
> 
> Resulttable :
>    a
> ------
> 10
> 12
> 12
> 12
> 14
> 14
> 15
> 
> I need the value which is repeated most in the resulttable (in this example
> 12 ) !
> 
> 
> Any suggestions ?
> 
> 
> Thx
> berger
>

Re: count different values in column

From
Fduch the Pravking
Date:
On Fri, Apr 19, 2002 at 04:40:35AM -0500, Tod McQuillin wrote:
> On Fri, 19 Apr 2002, Albrecht Berger wrote:
> 
> > Now I want to count the different values in the column.
> >
> > Resulttable :
> >    a
> > ------
> > 10
> > 12
> > 12
> > 12
> > 14
> > 14
> > 15
> >
> > I need the value which is repeated most in the resulttable (in this example
> > 12 ) !
> 
> SELECT a, count(a) FROM resulttable GROUP BY a ORDER BY 2 DESC

Note that count on NULL values will always give you 0:

fduch=> INSERT INTO resulttable VALUES (NULL);
INSERT 1702218 1
(repeat some times)

fduch=> SELECT a, count(a) FROM resulttable GROUP BY a ORDER BY 2 DESC;a  | count
----+-------12 |     314 |     210 |     115 |     1   |     0
(5 rows)

-- 
Fduch M. Pravking