>Dana,
>
>So WHY the need for low-level file-I/O?? . . . the input *could* be appended into a DBF defined with fixed, single-byte first character and remaining characters (79, I presume) as a single field (if more than 255 chars, use as many long fields as needed).
>Then, depending on the value in first fields SUBSTR commands could be used to fill the appropriate *real* (output) table(s) from the second field.
>
>Cheers,
>Jim N
>
You know, you're right. I should have seen that myself. This, of course, would be much quicker.

Dana