Thanks for a response.

My point was the data does need to be "transformed" into the proper source date layout for SQL Server. I believe this is the intended purpose of the CDATE function.

HOWEVER, through experimentation I've discovered that some of my dates are accepted as VFP date layout in text (i.e., mm/dd/yyyy hh:mm:ss). It seems that the transformation script pukes when I have an empty date in the text string, and CDATE seems to handle that to provide a null date.

Now I have loaded several thousand records, but the script fails because of "missing column delimiter". This is a pain. Where is the error handling? Do I need to write something into the script to go on to the next row?