Postgres Pro
Downloads
Home   >   mailing lists

Thread: Possible to do this in one query?

Possible to do this in one query?

From
"John Oakes"
Date:
I have a table with a column that contains fail codes.  These codes are 1-9.
I need to get a count of each of the codes.  Right now I have separate
queries:

select count(*) from tablename
where failcode = '1';

If I use GROUP BY it obviously doesn't get the count for codes that don't
exist.  I need to have 0 returned though so I have a listing of all 9 fail
codes and their total, even if it is zero.  I tried UNION like this:

select count(*) from tablename
where failcode = '1';
UNION
select count(*) from tablename
where failcode = '2';

etc, but it too excludes the codes that have a count of zero.  Right now I
have resorted to using 9 separate queries.  Is there a way to do this with
one?  Thank you!

John Oakes

Re: Possible to do this in one query?

From
Ian Barwick
Date:
On Tuesday 26 March 2002 19:22, John Oakes wrote:
> I have a table with a column that contains fail codes.  These codes are
> 1-9. I need to get a count of each of the codes.  Right now I have separate
> queries:
>
> select count(*) from tablename
> where failcode = '1';
>
> If I use GROUP BY it obviously doesn't get the count for codes that don't
> exist.  I need to have 0 returned though so I have a listing of all 9 fail
> codes and their total, even if it is zero.  I tried UNION like this:
>
> select count(*) from tablename
> where failcode = '1';
> UNION
> select count(*) from tablename
> where failcode = '2';
>
> etc, but it too excludes the codes that have a count of zero.  Right now I
> have resorted to using 9 separate queries.  Is there a way to do this with
> one?  Thank you!

If you have a seperate table containing all possible fail codes you can do 
something like this:

select fc.failcode, count(tn.failcode) 
from failcode_table fc
left join tablename tn 
on tn.failcode=fc.failcode
group by fc.failcode


Ian Barwick

Re: Possible to do this in one query?

From
"Andrew G. Hammond"
Date:
On Tue, Mar 26, 2002 at 01:22:41PM -0500, John Oakes wrote:
> I have a table with a column that contains fail codes.  These codes are 1-9.
> I need to get a count of each of the codes.  Right now I have separate
> queries:
>
> select count(*) from tablename
> where failcode = '1';
>
> If I use GROUP BY it obviously doesn't get the count for codes that don't
> exist.  I need to have 0 returned though so I have a listing of all 9 fail
> codes and their total, even if it is zero.  I tried UNION like this:
>
> select count(*) from tablename
> where failcode = '1';
> UNION
> select count(*) from tablename
> where failcode = '2';
>
> etc, but it too excludes the codes that have a count of zero.  Right now I
> have resorted to using 9 separate queries.  Is there a way to do this with
> one?  Thank you!

You need a table of failcodes first:

CREATE TABLE failcodes ( failcode INTEGER PRIMARY KEY,description TEXT
);

-- populate it:
INSERT INTO failcodes ( 1, 'Attacked by mob of angry kittens' );
-- ...
INSERT INTO failcodes ( 9, 'Wrong phase of moon' );

SELECT coalesce(count, 0) FROM ( SELECT failcode FROM failcodes NATURAL LEFT OUTER JOIN  SELECT count (*) FROM
tablenameGROUP BY failcode 
);

Or, be a sneaky bastard and add a fake record for each failcode, and
then count(*) - 1. :)

--
Andrew G. Hammond  mailto:drew@xyzzy.dhs.org  http://xyzzy.dhs.org/~drew/
56 2A 54 EF 19 C0 3B 43 72 69 5B E3 69 5B A1 1F              613-389-5481
5CD3 62B0 254B DEB1 86E0  8959 093E F70A B457 84B1
"To blow recursion you must first blow recur" -- me