Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!aramis.rutgers.edu!paul.rutgers.edu!birnbaum From: birnbaum@paul.rutgers.edu (Rich Birnbaum) Newsgroups: comp.databases Subject: Sybase Keywords: opinions, experiences, problems Message-ID: Date: 24 Jul 89 19:28:39 GMT Organization: Rutgers Univ., New Brunswick, N.J. Lines: 45 My company is in the early stages of developing a database application. We are using Sybase (version 3.0.2, going on 4.0 soon, we hope) and some quirks/bugs/problems have us a bit concerned. I'd be interested to know if anyone has come across any problems, specifically the ones listed below. I'd also welcome any other comments, pro or con, regarding Sybase. In particular, how you compare it to other client-server relational DBMS products (Ingres Version 6) and how it performs on 'large' tables (about 100,000 rows, with about 50 columns for a total of about 275 bytes per row). We will be joining such a table with a few smaller tables quite often in the application. 1) When running a complex query, the Data Server sometimes consumes the entire machine it is running on. By this I mean it uses close to 100% of the CPU, users already logged in to the Data Server get no response and new users cannot log into it. When this happens, we must kill the Data Server process and start it up again. 2) The following queries, to my surprise, returned different results. select * from REGISTRATION where not exists (select * from DD_1556 where REGISTRATION.FYNO_REG=FISCAL_YEAR and REGISTRATION.DOC_TYPE_REG=COST_CENTER_1556 and REGISTRATION.SERIAL_NO_REG=SERIAL_NUMBER) returns 0 rows, while select FYNO_REG,SERIAL_NO_REG from REGISTRATION where not exists (select * from DD_1556 where REGISTRATION.FYNO_REG=FISCAL_YEAR and REGISTRATION.DOC_TYPE_REG=COST_CENTER_1556 and REGISTRATION.SERIAL_NO_REG=SERIAL_NUMBER) returns 6 rows, my expected result. Note that the queries are identical except for the SELECT list. Please post responses or e-mail to either birnbaum@pica.army.mil or birnbaum@paul.rutgers.edu.