What about a special file containing all the information from the
sandbox files? Is that possible?
If the information being collected and keyed is identical to the
production files in every way, sure you could make one large "collection
file" of all needed info, but I would just mirror the normalized
production files as much as possible. Because most edits are applied to
combinations like <customer name> and <Customer number>, which I assume
in your case, has not yet been established.
I would make a bunch of Import-Files or Entry-Files that mirror the
formats of the production files.
When I create files this way, rather than confuse programs and myself
with library lists, just create new files and use the DDS keyword
FORMAT. This allows you to create identical fields and formats as the
original physical files, but the new physicals now get called by a new
name.
Then OVRDBF for edits that you can apply, and create new edits for when
you cannot apply an existing edit.
It is projects like this where you realize large data-editing programs
or routines are not good, and smaller, single-purpose edits are better,
where you address each data element (or combination of data elements) in
editing logic. Yes, it is a lot of functions to assemble and manage,
but in the long run, you realize that building editing logic should be
like building things with Legos (tm) building blocks.
- John V.
midrange.com
