Photomonkey wrote:
Ok, Just a wild guess. CVS = Comma-separated values.

Grin! I doubt users of such packages would agree :-) But I get your intent.

In English the decimal separator is a dot '.' In many languages the
decimal separator is a comma ','. I guess French uses a comma as the
separator. This is probably the most likely cause of the problem. In
general I think it is better that to "force" the programming logic to
use a specific Culture type both when reading and writing a cvs file.

Formatting of numbers and currency amounts is only part of a culture. My problem is specifically concerned with the extra characters that get added to a supposedly plain text file, that indicate UTF-8 was used to encode the file. These are :-

a) Three bytes inserted at the very start of the file, before the first 'real' byte of data. These are 0xEF 0xBB 0xBF.

b) Certain characters located in the top-half of the character set are not available in every code page. These are preceded by an extra byte to identify the particular code point i.e. you have a 16-bit character amongst the conventional 8-bit characters.

At present, only 3 characters cause me an immediate problem: namely, Alt-0188, Alt-0189 and Alt-0190 - the three 'fractional' symbols for a quarter, a half an three-quarters.

--
Regards,
Mike Fry
Johannesburg.

Reply via email to