Find files with non-printing characters (null bytes)

I hardly believe it, I might write an answer involving cat!

The characters you are observing are non-printable characters which are often written in Carret notation. The Caret notation of a character is a way to visualize non-printable characters. As mentioned in the OP, ^@ is the representation of NULL.

If your file has non-printable characters, you can visualize them using cat -vET:

-E, --show-ends: display $ at end of each line
-T, --show-tabs: display TAB characters as ^I
-v, --show-nonprinting: use ^ and M- notation, except for LFD and TAB

_{source: man cat}

I've added the -E and -T flag to it, to convert everything non-printable.

As grep will not output the non-printable characters itself in any form, you have to pipe its output to cat to see them. The following example shows all lines containing non-printable characters

Show all lines with non-printable characters:

$ grep -E '[^[:print:]]' --color=never file | cat -vET

Here, the ERE [^[:print:]] selects all non-printable characters.

Show all lines with NULL:

$ grep -Pa '\x00' --color=never file | cat -vET

Be aware that we need to make use of the Perl regular expressions here as they understand the hexadecimal and octal notation.

Various control characters can be written in C language style: \n matches a newline, \t a tab, \r a carriage return, \f a form feed, etc.

More generally, \nnn, where nnn is a string of three octal digits, matches the character whose native code point is nnn. You can easily run into trouble if you don't have exactly three digits. So always use three, or since Perl 5.14, you can use \o{...} to specify any number of octal digits.

Similarly, \xnn, where nn are hexadecimal digits, matches the character whose native ordinal is nn. Again, not using exactly two digits is a recipe for disaster, but you can use \x{...} to specify any number of hex digits.

_{source: Perl 5 version 26.1 documentation}

An example:

$ printf 'foo\012\011\011bar\014\010\012foobar\012\011\000\013\000car\012\011\011\011\012' > test.txt
$ cat test.txt
foo
                bar

foobar

        car

If we now use grep alone, we get the following:

$ grep -Pa '\x00' --color=never test.txt

        car

But piping it to cat allows us to visualize the control characters:

$ grep -Pa '\x00' --color=never test.txt | cat -vET
^I^@^K^@car$

Why --color=never: If your grep is tuned to have --color=auto or --color=always it will add extra control characters to be interpreted as color for the terminal. And this might confuse you by the content.

$ grep -Pa '\x00' --color=always test.txt | cat -vET
^I^[[01;31m^[[K^@^[[m^[[K^K^[[01;31m^[[K^@^[[m^[[Kcar$

sed can.

 sed -n '/\x0/ { s/\x0/<NUL>/g; p}' file

-n skips printing any output unless explicitly requested.
/\x0/ selects for only lines with null bytes.
{...} encapsulates multiple commands, so that they can be collectively applied always and only when the /\x0/ has detected a null on the line.
s/\x0/<NUL>/g; substitutes in a new, visible value for the null bytes. You could make it whatever you want - I used <NUL> as something both reasonably obvious and yet unlikely to occur otherwise. You should probably grep the file for it first to be sure the pattern doesn't exist before using it.
p; causes lines that have been edited (because they had a null byte) to show.

This basically makes sed an effective grep for nulls.