Custom drivers
If your operating system supports dynamic linking of libraries, you can use a custom driver to extend the functionality of the to support different file types or access mechanisms.
For example, you could implement a custom interface to load data from a structured file, a high-speed communications link, or another program that generates data to be stored in the database.
The onpload utility accesses the custom code through the driver name that you assign to the record-format definition. When onpload references a record format, the driver that the record format specifies is examined. If the driver name does not match one of the standard drivers (Fixed, COBOL, Delimited), onpload looks into the custom-driver function table to find the custom driver.
The custom-driver code reads data into buffers during a load job and writes out buffers during an unload job. By following the coding procedure discussed in this section, you can use the parallel I/O facilities of the HPL to manipulate data buffers, or you can use a custom driver to replace the HPL I/O facilities with your own I/O functionality.