I believe that there are other things to consider here. First, you mention using a joined SQL statement. Is this an actual statement within your code or is this using a stored procedure/view? If you use a stored procedure in place of an actual statement in code, your time should be improved with the SP. A view, depending on the complexity of the join, should be similar in performance if your DBMS "compiles/caches" your SPs/views.
You also need to consider the overall use of the database. Are you making updates? If so, are a lot of users making updates at the same time? If you plan to make updates to the dataset and then send the updates at one time per user using ADO.NET dataset update functionality, then the separate tables may be best regardless of performance. Otherwise, you can do whatever you want with the dataset and then build datacommand objects for your update statements. Christopher Reed Web Applications Supervisor Information Technology City of Lubbock [EMAIL PROTECTED] "The oxen are slow, but the earth is patient." >>> [EMAIL PROTECTED] >>> If I load 3 datatables into a strong typed dataset with relations and have the strong-typed dataset bound to a grid with lookups, it runs in 3.47 seconds. If I do the join in sql and bring back one datatable and have the grid bound directly to the strong typed dataset, it takes 3.35 seconds. This is with 31,000+ rows of data in the main table. I find it hard to believe it's doing table scans. =================================== This list is hosted by DevelopMentor® http://www.develop.com View archives and manage your subscription(s) at http://discuss.develop.com
