Sstrhsrtj

I have a text file which has content like this:

abc.tar^@xxx.tar^@yyy.tar^@ 

Say for example I have this content in a file named abc.txt and I want to split the content and write the first two entries into a new file.

(e.g), the new file would look like this:

abc.tar^@xxx.tar^@

Is there any command to perform this operation?

I'm guessing this question is related to that one, correct?

In that case, wouldn't replacing the '^@' with a newline be more worth your while?
In the following, I'm guessing you mean '^@', the ASCII NUL byte:

$ sed 's/o000/n/g' abc.txt | head -n 2

abc.tar

xxx.tar

So you need

sed 's/o000/n/g' abc.txt | head -n 2 > newfile.txt

Explanation

This substitutes a newline (n) for every NUL byte (o000) the o part means that what follows is a byte in octal notation. The output is then piped to head -n 2 which extracts the first two lines; and the resulting lines are redirected (>) to the file newfile.txt.

If it's important to you that the file names be separated by '^@', however, you can use this:

perl -nl000 -e '

    $num_lines =2 ;

    push @a,(split /00/)[0..$num_lines-1];

    print $_ for @a' abc.txt > newfile.txt

Replace the value of $num_lines above as needed to grab the first $num_lines lines from the file.

Explanation

• The -n switch tells perl to run the code on each line of the input file


• The -l000 sequence tells perl to set the output record separator
  (the character printed after every string) to the NUL byte (000).


• The -e switch tells perl that the string that follows is a code to execute.


• The split function splits each input line with the NUL byte as delimiter,takes the first $num_lines ([0..$num_lines-1]) results and puts them into the array @a. Notice that the "current input line" part is nowhere specified in the function call. This makes use of the fact that the default scalar variable in Perl ($_) is the default argument of the split function (among others) when no argument is supplied.


• The final foreach loop prints every element in @a (again note how $_ is the default iterator for the foreach loop). Since we have set the output record separator to octal 000, we get the results separated by the NUL byte as before.

Is this:

awk -F"@" '{print $1"@"$2"@"}' abc.txt > newfile.txt

good enough for you?

Try running:

sed -r -i 's/^(.*)@.*@.*$/1/' file

Here's an example using Perl:

$ perl -ne '@F = split(/@/,$_); print "$F[1]@$F[2]@";' abc.txt > newfile.txt

The above does the following:

• 
  @F = split(/@/,$_) - splits the contents of file abc.txt a line at a time up based on the character @ and stores the resulting chunks in an array (@F).


• 
  print "$F[1]@$F[2]@" - prints the first 2 columns, (1 & 2), from array @F and inserts a at sign (@) in between each column.

Awk can use any character as the record separator (with newline as the default), except that some implementations don't support the null byte as the separator. Gawk (GNU awk), the default awk on most non-embedded Linux installations, supports nulls.

gawk -v RS='' -v ORS='' 'NR <= 2 {print}'

This can be shortened to gawk -v RS='' -v ORS='' 'NR <= 2' since printing the record is the default action.

For a large file, you'd better quit after the second line.

gawk -v RS='' -v ORS='' 'NR==3 {exit} {print}'

Alternatively, you can use head. There's no option to use a null byte instead of a newline as the record separator, but you can swap the two characters, call head, and then swap back.

tr 'n' 'n' | head -n 2 | tr 'n' 'n'