Re: [GENERAL] Aggregates and Joins - Mailing list pgsql-general

Hello

If I have well understood what you want, you can do it by using a "group
by" clause. In the case you proposed, it should be used like:

select ref, sum(price)
from second_table
group by trans_id ;

I've tried this and it works for my database. Hope it helps


Bob Dusek wrote:
>
> Hello,
>
> I've kind of run into a wall, just now, and I don't know if I can climb it
> with Postgres.  In order to fully explain my problem, I've got to actually
> go into detail about my database setup... so, if you're willing to
> listen/read for a minute, I appreciate it.  Just so you know, my problem -
> as I see it - is related to the limitation of aggregate function use.
>
> In my database, I have two tables.  In one of the tables, I have a primary
> key 'trans_id'.  In the other table, I store vital information associated
> with 'trans_id'.  More specifically:
>
> table first_table (
>
>    trans_id int,  -- my primary key
>    otherfields blah
>    blah blah
>
> );
>
> table second_table (
>
>    this_id int,
>    price money,
>    ...
>    ..
>    trans_id int, -- my primary key association
>
> );
>
> In second_table, there may be any number (greater than zero) of records
> associated with a single record in first_table (ie. if I have a record in
> frist_table with a 'trans_id' = 33, I may have 1 or 100 records in
> second_table with a 'trans_id' = 33).
>
> What I need to do is to obtain the sum of the field 'price' from
> second_table for every distinct 'trans_id' (all of which can be found in
> first_table).  Right now, I do not know of any way to do this, other than
> by using my scripting language to first get all of the 'trans_id' values
> from first_table - "select trans_id from first_table" - and then (for
> each and every one of the records returned) initiating a separate query -
> "select sum(price) from second_table where trans_id = onetransid" - that
> would use the aggregate function sum() on the price field in second_table.
> This isn't a problem if I'm only trying to get the sum(price) for a
> hundred or so distinct 'trans_id' values.  However, I will (more often
> than not) be trying to do this for thousands of distinct 'trans_id'
> values.  Is there a quicker way to do this?  Is it possible to use the
> aggregate functions and have more than one row returned?
>
> Here's some pseudo-code to look at, if you'd like:
>
> // set up the query
> $query = "select trans_id from first_table";
> // execute the query and obtain a pointer to the results
> $result = dbExec($postgresConnection, $query);
>
> // find out how many records were returned
> $numrecords = pg_NumRows($result);
>
> // obtain the 'trans_id' values, one by one, and sum() the price
> for($i = 0; $i < $numrecords; $i++) {
>
>         $trans_id = pg_Result($result, $i, 0);  // record $i, index 0
>
>         $subquery = "select sum(price) from second_table " .
>                     "where trans_id = $trans_id";
>         $subresult = dbExec($postgresConnection, $subquery);
>
>         // only one record - to my knowledge - can be returned
>         $pricesum[$trans_id] = pg_Result($result, 0, 0);
>
> }
>
> Any help you can give will be graciously accepted.....
>
> Thanks,
>
> Bob

--
La risa es la actitud primigenia hacia la vida:
un modo de acercamiento que pervive sólo en
criminales y artistas
            -- Oscar Wilde