>>>I'm developing have a little utility that imports data from .CSV files. I have problems with some dates. I have set:
>>>
>>>SET DATE BRITISH
>>>SET CENTURY ON
>>>
>>>Dates come in british form (dd/mm/yyyy), they are NOT enclosed in quotation marks, they arrive to date type fields. After the APPEND FROM these dates suffer changes:
>>>
>>>CSV => VFP Cursor
>>>16/05/1968 => 16/05/2019
>>>19/06/2011 => 19/06/2020
>>>02/04/1973 => 04/02/1973
>>>11/08/2015 => 08/11/2015
>>>04/05/1938 => 05/04/1938
>>>26/07/2016 => 26/07/2020
>>>22/05/2004 => 22/05/2020
>>>24/03/2015 => 24/03/2020
>>>13/09/1980 => 13/09/2019
>>>16/05/1963 => 16/05/2019
>>>
>>>I tried changing to SET DATE AMERICAN and it becomes worst, the dates which previously are converted to end in 2020 appears blank.
>>>
>>>Does anyone have an idea?
>>
>>You might want to look at SET CENTURY TO xx ROLLOVER yy and SET STRICTDATE. But you're seeing some weird results, I'm not sure they would apply in your case.
>>
>>Some settings such as SET CENTURY are scoped per data session.
>>
>>If the above don't help then I think Thomas has the right idea, parse the input file, perhaps use an unambiguous date format such as {^2016-09-10}.
>
>
>Thank you for your advice, but I have no way to ask the csv provider to change the csv file.

There's no need for the provider to change the file, you can create a modified one for your own use i.e. parse the dates into something accepted unambiguously by VFP.