>The rollover only applies to ambiguous dates (dates without a century). The data looks like 6/15/1999. This is a known bug in the VFP 9.0 bug list that was supposedly corrected with SP2. Such is not the case with us. If we save the file as an Excel workbook, the append works fine. It's only a problem with CSV files.

Forgot that one because I always export CSV, don't need to import... but, just a thought: can't you import into a cursor which would have a c(10) field for dates and an extra field for each date? So when you append all these date fields come in as they are, and then you just

replace all dDateField with ctod(cDateField)  ;
  , dDateField2 with ctod(cDatefield2)...

and append into your target table from the cursor.

>>>CSV. But when I open the csv file with Excel, that's the format the date fields are in.
>>
>>So Excel recognizes these strings as dates, and interprets (and displays) them so. What's the actual content of your csv file, what are your set("century"), rollover settings?