Re: READE implementation difficulties. RE: When is %EOF not an %EOF

> Without the need for the database engine
> to read and lock the next sequential record,
> couldnt READE be implemented solely in
> the compiled program code?

I'd need to do a READP in order to emulate READE in code.  Imagine this key
list:

Cust-Item
1-1
1-2
1-3
2-1
2-2

In a READE loop, limited to a partial key of CUST 1, when the runtime reads
customer 2, it gives me EOF and leaves the database fields in the program
intact.  That is, the program has the values from the last item of customer
1.

If I were to try that with READ/compare, when I READ customer 2, I now have
the values from the first item of customer 2.  I can set my internal EOF,
but I have the 'wrong' values and now need to do a READP to get them back.

Of course, the business rules may not give a hoot about whether the values
have been modified on our pseudo EOF, so that may be a non-issue for those
situations.
  --buck