On Wed, Jan 26, 2005 at 11:22:48AM +0100, Thomas Chille wrote: > In the meanwhile i gave every record an position counter (per tour and > employee), derrived from begin_time, per trigger, and merged them in > this way:
How are you generating the position counter? A sequence? If you multiply by some factor it might eliminate the need to re-number everything. > SELECT * FROM work LEFT JOIN plan USING(id_tour, id_employee, counter) > UNION > SELECT * FROM work RIGHT JOIN plan USING(id_tour, id_employee, counter) > > Thats works for me but if the tourleader change one worktime record > the counters have to be recalculate and the order of the records will > change. > > This is hard to handle and i think i will do all merging > programmatically in the app or with an ppgsql function. > > > BTW, I've never seen the convention id_employee; people generally use > > employee_id. Is it more important to know that you're talking about an > > ID or that you're talking about an employee? Just food for thought. > > I'm not the father of this strange naming convention :) I hate getting stuck with other people's stupidity. :) -- Jim C. Nasby, Database Consultant [EMAIL PROTECTED] Give your computer some brain candy! www.distributed.net Team #1828 Windows: "Where do you want to go today?" Linux: "Where do you want to go tomorrow?" FreeBSD: "Are you guys coming, or what?" ---------------------------(end of broadcast)--------------------------- TIP 4: Don't 'kill -9' the postmaster