Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!lsuc!dave From: dave@lsuc.UUCP Newsgroups: comp.databases Subject: Re: How good is Empress? Message-ID: <1661@lsuc.UUCP> Date: Tue, 24-Mar-87 21:00:29 EST Article-I.D.: lsuc.1661 Posted: Tue Mar 24 21:00:29 1987 Date-Received: Wed, 25-Mar-87 07:21:18 EST References: <449@agsm.unsw.oz> <349@hscfvax.UUCP> Sender: root@lsuc.UUCP Reply-To: dave@lsuc.UUCP (David Sherman) Organization: Law Society of Upper Canada, Toronto Lines: 41 Summary: well designed but sometimes slow In article <349@hscfvax.UUCP> pavlov@hscfvax.UUCP (840033@G.Pavlov) writes: >In article <449@agsm.unsw.oz>, bobm@agsm.unsw.oz writes: >> >> We are considering acquiring a database language to operate on a >> Vax 11/780 under System 5.2.2. ..... >> but have rarely seen reference to Mistress/Empress, which has >> been highly recommended. What experience have users had with Empress? >> How does it stack up against the better known databases (Ingres, Informix), > > Feature-wise, Mistress/Empress compares well. It also has the attribute that > one can easily imbed database manipulation commands in a shell script (tho > it is not difficult to write a simple one to do the same in others; I did > that for our Ingres installation). > > I would, however, see if the company has worked on the speed of join operat- > ions .... I agree with the above. We run Mistress/Empress on two machines, one for a substantial student records system (2,200 students, 80 fields of information per student) and one for educational program registrations and publications sales (~15,000 of each over a year). Overall, the design features are excellent. However, retreivals, particularly when joining across tables, are painfully slow. I rewrote one nightly-statistics program a couple of weeks ago. Using the report write (4GL) and joining across tables, it was taking 5 hours; rewritten in C to load the program names and number into arrays in memory instead, it now takes 17 minutes. If you're short on memory, note that Empress is also kind of big (300K on a 1Mb system makes it a bit slow when a number of users start making queries at the same time). I haven't compared it to other DBMSs in this respect, though. David Sherman The Law Society of Upper Canada Toronto -- { seismo!mnetor cbosgd!utgpu watmath decvax!utcsri ihnp4!utzoo } !lsuc!dave