This is not something I have seen specifically although there is always the possibility that something in the files is causing them to be interpreted incorrectly. Formatting features and printer control character come to mind here.
And of course a file that is named for and looks like a text file may not be a text file in reality.
Without seeing the files it is difficult to be more helpful. You could check the files using notepad or wordpad to see what they look like when editing. Or you could use the Datawatch Examine.exe utility to report the contents on a character by character basis.
What are the sources for the files? Is there a special process run to obtain them?
If the data are not confidential and you are able to share the problem examples I would be happy to take a look at the files to see if I can spot anything and suggest a way to work around the problem.
First of all thanks for answering.
The txt file contains downloaded data from a SAP system. Opening it witj Notepad makes it appear a good textfile.
I have used the examine.exe but cannot read anything from the result. I wanted to add a file to this post but the feature is not enabled....
Another strange thing is that when I export the data to a dbf file, the total of lines exported is not always the same, e.g. one time there were 276'753 lines, another time 276042 lines. But each time the same source file.
Since the source textfile is around 150 MBs, I am wondering if there is any kind of size issue on Notepad side.
Could I mail you the screenshots I made or an extract of the file (the entire file itself is confidental).