Why does Linux report the following as a binary file?

Question

The file is here

http://dl.dropbox.com/u/12337149/history.csv

It views just fine in Notepad but Centos complains about the following

"history.csv" may be a binary file. See it anyway?

If I choose yes the contents are not properly displayed

Answer 1

http://www.thelinuxblog.com/working-with-csv-files-in-bash/

Answer 2

What are you trying to do in Centos that is complaining? Just looking at the file, the only thing that stands out to me is the UTF-16 Byte Order Mark. I'd guess that whatever is opening the file is not dealing w/ the character set correctly and instead thinks it is a binary format.

I have come to this conclusion by doing: file history.csv and getting

history.csv: Little-endian UTF-16 Unicode text, with CRLF, CR line terminators

In the shell, when I cat the file I get

��Type Summary Date Ref Description Period Opening Ccy Size Closing Amount
WITH Commissions 14/04/11 4553585 Aurelian Oil & Gas Plc COMM G42H4S "-" - � - 0 -10.16

Then I looked at the hex values of the unknown characters (assuming that shows above) and saw:

0000000: fffe 5400 7900 7000 6500 0900 5300 7500 ..T.y.p.e...S.u.

Looking at the UTF-16 entry on wikipedia, It says that is the Byte order mark, showing the endienness.

So it was a very long path to get to "don't know...maybe it's not dealing with the character set" :)