We update a few fields in every record on a file with over 13 million records. Masking a copy of production data on a performance test environment. It is run on a two processor 595 box under OS7.1. The job runs for about 36 hours. If I wrote an RPG program, I would open the file non-keyed to process sequentially and, I assume, it would update all records in the I/O page making the I/O process more efficient. Can I make SQL do the same or do I need to write an RPG program to see if it runs quicker? The record length is 980 and it's a DDS file, before anyone suggests I can't convert to SQL as it's a package and I am limited in what I can do.
Software/Hardware used: two processor 595 box under OS7.1