Split databases and Cursors? (1 Viewer)

[tmaccabe]

I'm having trouble understanding which cursors to use when developing for a split environment, i.e. a local mdb linked to a remote mdb on a network share.

Does a split architecture imply that I should always use a client-side cursor?

What happens when I use each cursor in this architecture? Where does the processing occur? How does it effect network traffic?

Can you recommend a good reference? I've been trying to use OLE DB as much as possible, and therefore ADO. I'm also looking for a best practices reference that speaks to this. We're running Access 2000, but will soon upgrade to Access 2003.

Thanks, -Tony.

 

[The_Doc_Man]

The cursors always exist where the query is executed.

If you use Access as the front-end for Access queries, the cursors exist on your workstation. Your actions to navigate the result use data stored in Access to find last/next/previous/first record.

If you use a pass-through query and the SQL engine on the other end of the ODBC executes the SQL, the cursor is on the server side.

Server-side SQL is far preferable in terms of network traffic. However, if both the local and the shared DBs are .MDB files from Access, you are always using client-side cursor because the network share is a FILE SERVER ONLY. No SQL at all. In this case, local tables are more efficient from the network point of view but less convenient for multi-user data sharing issues.

 

[tmaccabe]

Thanks! I switched over to DAO as the OLE DB thing was a red herring. Now I don't have to worry about cursors as you don't define in DAO.