Date: Sat, 29 Dec 2001 16:20:35 +0100
From: QdlatY <qdlaty@wielun.dhs.org>
Reply-To: QdlatY <qdlaty@wielun.dhs.org>
Message-ID: <1132751000.20011229162035@wielun.dhs.org>
To: pgsql-sql@postgresql.org
Subject: difficult query (for me)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hello

I will try to explain my problem as clearly as i can.

So, I have tree tables
Named:

Tables
Threads
Posts

It is typicall structure of discussion phorum.
Now, Tables has a column ID (unique random index)
Threads has a columns ID (index) and TableID (index of one row of tables 
table)
Posts has attribute ThreadId and CreateTime

(i only wrote important columns to this query)

Now, I have Tables row with ID = 10
And then i want to have list of all Threads (belongs to Table with id = 10, 
so WHERE Threads.TableId = 10) sorted by CreateTime of newest Posts row 
belong to every thread.

F.E.
Table.iD = 10

Thread
ID TableID
1  10
2  10
3  11

Posts
ThreadId  CreateTime (i will use nubers only to explain)
1         10
1         20
1         25
2         43
3         02


So i want to have result from Table ID = 10:
Threads
ID
2    (because  CreateTime of Posts from 2 Thread is highest)
1    (and so on...)

(Thread 3 not belongs to Table.ID = 10)

Of course result can contains adequate Posts columns but don't must to...

I hope you will able to help me, i think it's not difficult but i'm a 
little newbie, so, i gave up on this :PPP

Best regards
QdlatY