An rdoConnection is a connection to one ODBC data source that uses the user name and password of the rdoEnvironment in which the rdoConnection resides. You can set properties on an rdoConnection to specify what kind of cursor it uses (server-side, client-side, or none) and also the Login and Query time-outs.
The cursor type specifies where rows of data should be buffered to allow the user to scroll forward and backward through the resultset. A server-side cursor puts more of the burden on the server and the network. Only the data the client requires at any one time is transferred across the network; generally, this means just enough data to present on one screen. Using a server-side cursor usually will reduce network traffic, but it can also increase network traffic under certain circumstances, such as when data is passed from the server to the client many times as the user scrolls up and down the resultset.
Using a client-side cursor means that the complete resultset is transferred to the client at one time and then is held in memory or on disk. A client-side cursor provides a more predictable load on the whole system, but it can slow everything down as large amounts of data are transferred across the network to be stored temporarily on the client PC. This time is wasted if all the user then does is pick the first record.
Using no cursor is a very quick method of fetching rows, but it means the client program can read the resultset only once and can’t move backward through the data. It’s therefore a useful method when populating static list boxes and combo boxes and when generating reports. It’s also useful in three-tier applications when the middle tier makes a quick connection to the back-end data server to retrieve information in one large chunk, which it then passes on to the front-end presentation application in smaller chunks.