RE: Read stream files by data/time?

This is part of a larger process that will involve DB2 I/O and 3rd party API calls. I didn't want to make the removal of duplicates a completely separate step.

The file names are 28 characters long. The first 8 are the "primary key", and the remaining 20 are complete garbage. For example, these files are duplicates. The newest is binding, and the remaining are obsolete.
12345678aaaaaaaaaaaaaaaaaaaa.pdf
12345678bbbbbbbbbbbbbbbbbbbb.pdf
12345678cccccccccccccccccccc.pdf


Jim Franz,
That's a solid plan. Read the dir in natural order using RPG, use the stat() API to get the date/time and write a cross-ref table. Then use that table, ordered by date/time, for my process.

Alan Campin,
I do all 3. In those languages, you'd read the dir contents into an array and then sort the array. If an array of that size didn't melt my PC down to slag, it would definitely make it unusable for hours.

Roger Harman,
It's not clear exactly how that app determines duplicates. I'm also skeptical whether my PC can handle that big of a dir.

John Yeung,
Native Java would be an option. I don't know Python, and I doubt it's even installed on my i. I tend to stick with RPG because it's the only language the other devs know.

Bryan Dietz,
Wow, that doc is beyond my "Dick and Jane"-level C skills.