Hello Jeroen...

In my case, when and how I connect/disconnect is based on the number of records, whether or not the data is readonly or read-write...etc. If the remote table is a primary data store that I will be reading and writing to and I expect a massive number of records, I like to open it up -once- on application startup and close it on application shutdown.

There should be few implications from the server side of keeping the connection open.

Each time you connect to a table, there's a bit of a lag. If you connect and disconnect for each command, that lag may become intolerable.

Also consider playing around with the FetchAsNeeded, MaxRecords, and FetchSize properties. These can have dramatic effects in how "clean" your application runs and the responsiveness.


>When should you use a connect/disconnect ?
>At the beginning and at the end of your application or before and after every command ? What are the results towards the server of keeping a connection open like that ? Any comments would be appreciated. I personally think you should connect and disconnect after each command/command group. I'm creating an application with remote views to simulate the real client/server environment and I want to know what the effects will be when I move to a real Oracle server/Sql Server situation.