RE: Questionable practice in file handling in CL

Duane,

I agree with the method you have outlined, as I have had experience within the last 8 years of doing that same idea in order to apply a vendor update. However, the file had way more than 20,000 to 30,000 rows in it. As I recall, it was in the millions or possibly billions of rows. The runtime saved was in terms of hours, and I am thinking it was at least 24; it was substantial. Each row in the file was touched by the vendor update process, so each LF index was "touched". Deleting the LFs, running the update over just the PF, and then recreating the LFs got the update down to about 6 hours [including the LF rebuild] or so, which was significant, since we were thinking we could not upgrade over the weekend because the rebuild took so long. That was on a V5R4 machine at the time.

Removing the LFs did greatly improve the processing time, but the file did have a ton of records in it [it was a sales tax history file].

Doug