>Here is the idea behind it:
>-Analyze text file to determine its type
>-Create a cursor that'd hold all the information as possible from text file (ie: if it's determined to be a tab delimited file, tab count gives generic intermediate cursor's field count. Start with all of them 254 in width, append then shrink widhts)
>-Map user columns to target's columns
>-Rename columns that'd be imported
>-Append from (or scan...insert...endscan)
>Roughly that's it. Our users were importing various formats from mainframes and this was the one I came up with.
>Cetin

Cetin

I totally trust you on importing matters! You proved it over my CIF import data. So if you've said this is the way to go about it I know it must be the optimum. And it sounds fairly simple. It looks like this is way to go.

Thanks

Terry