I want to search for the match count of a specific word/pattern in all files below a certain directory. However, I want the number to appear directly after the filename.
I did the following but it writes the count to a new line instead of the last line
grep -rc "WORD" /home/user/data >> file
Actual output
foo
12
Desired output:
foo12
With >> file, the file is opened on stdout in append mode. So any output goes at the end of the file.
grep writes 12\n. Your file contained foo\n beforehand. So you get foo\n12\n. To get foo12\n, you'd need not to output at the end, but at the last byte before then end, to overwrite that \n with 12\n.
Some shells have operators for that.
With zsh:
zmodload zsh/system
if sysseek -u 1 -w end -1; then
grep -c WORD /home/user/data
else
syserror -p "Can't seek to just before the end: "
fi 1<> file
With ksh93:
grep -c WORD /home/user/data 1<> file >#((EOF - 1))
<> is the redirection operator that opens files in read+write mode. Here, we don't care about the read part, but we use <> as contrary to >, it doesn't truncate files.
In any case, that is not an append mode. You'll see a difference if some other process also writes to the file or truncates it at the same time (unlikely to be a problem in your case).
Another approach could be to truncate the file to remove the trailing \n, and then open the file in append mode.
With GNU truncate:
truncate -s -1 file && grep -c WORD /home/user/data >> file
You can use sed to append at the end of the last line:
sed -i '$ s/$/'$(grep -rc "WORD" /home/user/data)'/' file
See also: append text with echo without new line
