Mime-Version: 1.0
Message-Id: <6734364.1181577779@[4.42.179.151]>
In-reply-to: <20020828161241.50a8512d.gry@ll.mit.edu>
Date: Wed, 28 Aug 2002 14:25:17 -0700
To: pgsql-sql@postgresql.org
From: eric soroos <eric-psql@soroos.net>
Subject: Re: select question
Content-Type: text/plain; charset=us-ascii

On Wed, 28 Aug 2002 16:12:41 -0400 in message <20020828161241.50a8512d.gry@ll.mit.edu>, george young <gry@ll.mit.edu> wrote:
> [postgreql 7.2, linux]
> I have a table T with columns run, wafer, and test:
>    T(run text, wafer int, test text)
> Given a run and a set of wafers, I need the set of tests that match
> *all* the specified wafers:
> 
> run	wafer	test
> a	1	foo
> a	2	foo
> a	3	foo
> a	3	bar
> 
> E.g.
>   Given run 'a' and wafers (1,3) I should get one row: foo, since only foo matches both 1 and 3.
>   Given run 'a' and wafers (3) I should get two rows: foo,bar, since both foo and bar match 3.
> 
> Is there some neat way to do this in a single query?
> 

select test from T
  where run='a' 
    and wafers in ('1','3')
  group by test
     
eric