>>>Our client(s) is send us CSV files with demographics of people they serve. The problem is that a few of the rows in the CSV have names and/or addresses with embedded comma in them. This is throwing the APPEND a problem. The wrong data is ending up in the wrong columns.
>>>Can anyone suggest how to prevent this from happen at my end.
>>
>>Perhaps you should process such files line by line?
>
>Yes, I have considered that. But there are 10s of thousands of rows, and several files of them. The performance hit would be great.

One other possibility, I think, is to have a file with several extra dummy columns. If you import the file using APPEND from command and find that few extra columns got populated, then you know you'll need to re-process that file.