Обсуждение: [GENERAL] Help with query (fwd)

---------- Forwarded message ----------
Date: Fri, 5 Nov 1999 09:35:25 -0400 (VET)
From: Vegeta <vegeta@cuaima.ica.luz.ve>
To: pgsql-general@postgresql.org
Subject: Help with query

Dear list members:

I am new to the list and to SQL. I am using PostgreSQL 6.4.2 on Linux and
want to do the following query:
Given a table "works" with fields "idworker" (int4) and "timeofwork"
(datetime),  i want to have a query who has the number of hours worked by
each "worker" in the form:
 idworker   hoursworked

I tried with:
SELECT idworker, COUNT(DISTINCT int8(timeofwork)/3600) as hoursworked
FROM works
GROUP BY idworker

but pgsql does not accept the DISTINCT inside the count (I read in a book
that there are some RDBMS that do accept it).
Is there another way to acomplish this?
Do I need to create a user-defined aggregate? If so, how should I do it?

Thanks in advance,
Guido Urdaneta

PS: Sorry for my bad english

On 1999-11-05, Vegeta mentioned:

> Given a table "works" with fields "idworker" (int4) and "timeofwork"
> (datetime),  i want to have a query who has the number of hours worked by
> each "worker" in the form:
>  idworker   hoursworked
>
> I tried with:
> SELECT idworker, COUNT(DISTINCT int8(timeofwork)/3600) as hoursworked
> FROM works
> GROUP BY idworker

That query doesn't make much sense to me. I think what you meant was
something like this:

SELECT idworker, SUM(int8(timeofwork)/3600) FROM works GROUP BY idworker

I'm not sure about the appropriateness of your attempted typecast there,
though.

The timeofwork column should most likely be a timespan field. You can then
fill it with values like '5 hours 45 minutes'. Then your query could be:

SELECT idworker, SUM(date_part('epoch', timeofwork) * 3600) FROM works
GROUP BY idworker

--
Peter Eisentraut                  Sernanders vaeg 10:115
peter_e@gmx.net                   75262 Uppsala
http://yi.org/peter-e/            Sweden