Programming the Perl DBI

Driver Handles

Driver handles represent loaded drivers and are created when the driver is loaded and initialized by the DBI. There is exactly one driver handle per loaded driver. Initially, the driver handle is the only contact the DBI has with the driver, and at this stage, no contact has been made with any database through that driver.

The only two significant methods available through the driver handle are data_sources() , to enumerate what can be connected to, and connect() , to actually make a connection. These methods are more commonly invoked as DBI class methods, however, which we will discuss in more detail later in this chapter.

Since a driver handle completely encapsulates a driver, there’s no reason why multiple drivers can’t be simultaneously loaded. This is part of what makes the DBI such a powerful interface.

For example, if a programmer is tasked with the job of transferring data from an Oracle database to an Informix database, it is possible to write a single DBI program that connects simultaneously to both databases and simply passes the data backwards and forwards as needed. In this case, two driver handles would be created, one for Oracle and one for Informix. No problems arise from this situation, since each driver handle is a completely separate Perl object.

Within the DBI specification, a driver handle is usually referred to as $drh .

Driver handles should not normally be referenced within your programs. The actual instantiation of driver handles happens ``under ...