When I type this in Ubuntu:
# cat -A my.config
It outputs:
^@^@^@^@^@^@^@^@^@^@^@
But for:
# cat my.config
nothing shows up in the terminal.
So from bash, how can you tell if a file contains those strange characters?
Thanks!
3 Answers
Those characters are NULL bytes. You can search for them on a file with:
grep -Pa '\x00' /folder/file
You may also search for all non-ASCII characters on a file with this:
grep -Pa "[\x80-\xFF]" /folder/file
Finally, you could check for all non-ASCII characters on all files in a folder with the following code:
grep -Pa -r "[\x80-\xFF]" /folder
Note: the core of this answer relies on grep -P, which might not be available on all grep versions. Also, from grep man about the -P option:
-P, --perl-regexp Interpret PATTERN as a Perl regular expression. This is highly experimental and grep -P may warn of unimplemented features.
3 Comments
Charles Duffy
grumbles about use of GNUisms without explicit caveat or warning
Jamil Said
@Charles: Is it better now? If you think another caveat should be added please comment again.
Charles Duffy
Yup. (I was trying to decide whether to bring this up at all, since the OP explicitly specifies "Linux" in the title, and GNU grep is most common there -- but then, there are quite a lot of embedded Linux systems running Busybox, or a few running GNU grep compiled without the optional -- meaning, not-always-compiled-in even with latest GNU grep -- PCRE support).
Assuming "viewable" class in your post matches grep "graph" class, this should work
grep -la "[^[:graph:]]"
1 Comment
Charles Duffy
There are actually stronger guarantees that this will be available than the GNU-specific documentation you linked to -- the
graph character class is part of POSIX BRE, so absent -a this answer would be guaranteed to work on all POSIX platforms.Octal dump is your friend:
od -c my.config
filecommand, it will tell you if the file is a text file or something else.