Re: [SQL] Function Parameters in GROUP BY clause cause errors

When I use a parameter in a query which aggregates it fails with a GROUP BY error. What syntax can I use to avoid this error?

CREATE TABLE test (email_creation_datetime timestamp);
INSERT INTO test VALUES ('2006-03-20 09:00');
INSERT INTO test VALUES ('2006-03-20 09:15');
INSERT INTO test VALUES ('2006-03-20 09:30');
INSERT INTO test VALUES ('2006-03-20 09:45');

Query without parameters works fine:

select to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/30) * 30, 'HH24:MI'), 'HH24:MI') as TheInterval
from test em
group by to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/30) * 30, 'HH24:MI') , 'HH24:MI')

theinterval
09:30
09:00

But the same query with a parameter returns a GROUP BY error:

CREATE OR REPLACE FUNCTION EmailByInterval(IntervalMinutes int) RETURNS SETOF test AS $$
DECLARE rec RECORD;
BEGIN
FOR rec IN
select to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/IntervalMinutes) * IntervalMinutes, 'HH24:MI'), 'HH24:MI') as TheInterval
from test em
group by to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/IntervalMinutes) * IntervalMinutes, 'HH24:MI') , 'HH24:MI')
LOOP
RETURN NEXT rec;
END LOOP;
RETURN;
END;
$$ LANGUAGE plpgsql;

Query returned successfully with no result in 70 ms.

select * from emailbyinterval(30);

ERROR: column "em.email_creation_datetime" must appear in the GROUP BY clause or be used in an aggregate function
CONTEXT: SQL statement " select to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/ $1 ) * $2 , 'HH24:MI'), 'HH24:MI') as TheInterval from test em group by to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/ $3 ) * $4 , 'HH24:MI') , 'HH24:MI')"
PL/pgSQL function "emailbyinterval" line 3 at for over select rows

"Davidson, Robert" <robdavid(at)amazon(dot)com> writes:
> ERROR: column "em.email_creation_datetime" must appear in the GROUP BY =
> clause or be used in an aggregate function
> CONTEXT: SQL statement " select to_char(to_timestamp(EXTRACT(HOUR FROM =
> em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM =
> em.email_creation_datetime)::integer/ $1 ) * $2 , 'HH24:MI'), =
> 'HH24:MI') as TheInterval from test em group by =
> to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime) || =
> ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/ $3 ) * =
> $4 , 'HH24:MI') , 'HH24:MI')"
> PL/pgSQL function "emailbyinterval" line 3 at for over select rows

Hmm, this seems like a plpgsql deficiency. It feels it can generate a
separate parameter symbol ($n) for each occurrence of each variable it
passes into a SQL query. But for this query to be legal, the two
instances of IntervalMinutes have to be represented by the *same*
parameter symbol (notice they are not in the regurgitated query).

It would be more efficient anyway to not generate multiple parameters
for the same value, so we oughta fix this.

In the short run, the only workaround I can think of for you is to run
the query using EXECUTE.

regards, tom lane

Just Put aggregate function to the fields you selected.

Like this:

select to_char(to_timestamp(EXTRACT(HOUR FROM
max(em.email_creation_datetime)) || ':' || (EXTRACT(MINUTE FROM
max(em.email_creation_datetime))::integer/30) * 30, 'HH24:MI'), 'HH24:MI')
as TheInterval

from test em

group by to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime)
|| ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/30) *
30, 'HH24:MI') , 'HH24:MI')

_____

From: pgsql-sql-owner(at)postgresql(dot)org [mailto:pgsql-sql-owner(at)postgresql(dot)org]
On Behalf Of Davidson, Robert
Sent: Wednesday, March 22, 2006 1:12 PM
To: pgsql-sql(at)postgresql(dot)org
Subject: [SQL] Function Parameters in GROUP BY clause cause errors

When I use a parameter in a query which aggregates it fails with a GROUP BY
error. What syntax can I use to avoid this error?

CREATE TABLE test (email_creation_datetime timestamp);

INSERT INTO test VALUES ('2006-03-20 09:00');

INSERT INTO test VALUES ('2006-03-20 09:15');

INSERT INTO test VALUES ('2006-03-20 09:30');

INSERT INTO test VALUES ('2006-03-20 09:45');

Query without parameters works fine:

select to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime) ||
':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/30) * 30,
'HH24:MI'), 'HH24:MI') as TheInterval

from test em

group by to_char(to_timestamp(EXTRACT(HOUR FROM em.email_creation_datetime)
|| ':' || (EXTRACT(MINUTE FROM em.email_creation_datetime)::integer/30) *
30, 'HH24:MI') , 'HH24:MI')

theinterval

09:30

09:00

But the same query with a parameter returns a GROUP BY error:

CREATE OR REPLACE FUNCTION EmailByInterval(IntervalMinutes int) RETURNS
SETOF test AS $$

DECLARE rec RECORD;

BEGIN

FOR rec IN

select to_char(to_timestamp(EXTRACT(HOUR FROM
em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM
em.email_creation_datetime)::integer/IntervalMinutes) * IntervalMinutes,
'HH24:MI'), 'HH24:MI') as TheInterval

from test em

group by to_char(to_timestamp(EXTRACT(HOUR FROM
em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM
em.email_creation_datetime)::integer/IntervalMinutes) * IntervalMinutes,
'HH24:MI') , 'HH24:MI')

LOOP

RETURN NEXT rec;

END LOOP;

RETURN;

END;

$$ LANGUAGE plpgsql;

Query returned successfully with no result in 70 ms.

select * from emailbyinterval(30);

ERROR: column "em.email_creation_datetime" must appear in the GROUP BY
clause or be used in an aggregate function

CONTEXT: SQL statement " select to_char(to_timestamp(EXTRACT(HOUR FROM
em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM
em.email_creation_datetime)::integer/ $1 ) * $2 , 'HH24:MI'), 'HH24:MI') as
TheInterval from test em group by to_char(to_timestamp(EXTRACT(HOUR FROM
em.email_creation_datetime) || ':' || (EXTRACT(MINUTE FROM
em.email_creation_datetime)::integer/ $3 ) * $4 , 'HH24:MI') , 'HH24:MI')"

PL/pgSQL function "emailbyinterval" line 3 at for over select rows

I choose Polesoft Lockspam to fight spam, and you?
http://www.polesoft.com/refer.html

I wrote:
> Hmm, this seems like a plpgsql deficiency. It feels it can generate a
> separate parameter symbol ($n) for each occurrence of each variable it
> passes into a SQL query. But for this query to be legal, the two
> instances of IntervalMinutes have to be represented by the *same*
> parameter symbol (notice they are not in the regurgitated query).

> It would be more efficient anyway to not generate multiple parameters
> for the same value, so we oughta fix this.

Patch applied to HEAD and 8.1 branches.

regards, tom lane