>Steve,
>
>For this first client, I don't think know if we are going to have the time to test out both methods unless we go a real quick and dirty test.
>In terms of the next release of the app, then yes, I can see writing the code and creating enough test data to conduct a meaningful test.
>

I think a hard-coded test wouldn't be much faster than a (well-designed) generalized solution, because the bottlenecks are the disk hits, so a "quick and dirty test" would tell you a lot. But I understand the time constraint. :)

>Regardless of which way is ultimately 'faster' beginning to end in performance... I do concur with one observation that you made in regards to memos. We have an app that creates an ASCII file with EDI data (maybe 100k).
>We * used * to create a memo record and concatanate the data to the one record. Then create the ASCII file from the one memo record.
>We changed the app to do 'fwrite' and the runtime decreased to a small fraction of the time.
>

The method I was proposing is similar to how I currently create my EDI files (and ASCII files to be transferred between different apps).

>Once again, sincere thanks for your valuable input.
>

I'm glad to help...