Thanks for the input Perry,

I had originally written the code to make more extensive use of Fox's built in string manipulation functions, but the ambiguous nature of this data format makes it difficult to use the AT() and SUBSTR() functions effectively. The problem is that either "~" or "|" can be a delimeter, but can also appear in the data as an escaped character - "/~" or "/|".

An example of a key name and data chunk would be "|FIRSTNAME~MATT|". But "|FIRSTNAME~MA/|TT" is also valid.

So you can see my problems with using AT() and SUBSTR() to find the delimiters.

Aside from all that, can Foxpro not step through a large character string efficiently? Processing a 15k string takes 25 seconds (on a pretty slow machine). Doing the same in Java takes <1 second, and Java ain't no speed demon either.

In any case, I'm considering doing what you recommend and putting some more AT's and SUBSTR's.

Thanks again.

Matt