xargs usage syntax

Question

I am trying to search for a particular string within zipped files but cannot get the 'xargs' syntax correct.

The files unzip/zip correctly but xargs is actually searching for nothing (we are looking for failed TLS EMails).

Can anyone give me some pointers about the correct xargs syntax?

for filename in $( ls -1 ${HOST}-mail-2018${1}[0-9][0-9]* )
do
  filetype=${filename##*.}
  case $filetype in
    bz2)
        unzipper="bzip2 -d "
        zipper="bzip2"
        unzfile=${filename%.${filetype}}
        ;;
    gz)
        unzipper="gzip -d "
        zipper="gzip "
        unzfile=${filename%.${filetype}}
        ;;
    xz)
        unzipper="xz -d "
        zipper="xz "
        unzfile=${filename%.${filetype}}
       ;;
    *)
        echo "Unknown compression type for file $filename"
        break
        ;;
   esac
        #  Testing:    echo $unzipper $zipper $unzfile
        echo $unzipper $zipper $filename $unzfile

   eval ${unzipper} ${filename}
   grep 'Cannot .*TLS' ${unzfile} | sed 's/^.*]: //' | sed 's/:.*//' |  xargs fgrep
   eval ${zipper} ${unzfile}
done
exit 0

Answer 1

for filename in "$HOST-mail-2018$1"[0-9][0-9]*; do
    case $filename in
        *.gz)  g=zgrep  ;;
        *.bz2) g=bzgrep ;;
        *.xz)  g=xzgrep ;;
        *) printf 'Unknown filetype for "%s"\n' "$filename" >&2
           exit 1   # or continue or break
    esac

    "$g" 'Cannot .*TLS' "$filename"
done

Each compression tool comes with a corresponding grep tool. These are, for gzip, bzip2 and xz, called zgrep, bzgrep and xzgrep respectively. By using these there is no need to explicitly uncompress and recompress the files.

The script above picks the correct grep depending on the file suffix. One could arguably use plain grep for any unknown suffix. See below for how to do this without looking at filename suffixes (using the file tool).

Note how we don't need to use ls to loop over the set of files, and that the variable expansions need to be double quoted.

I ignored the sed calls that you have in your code as I don't know what the purpose of these are. I also removed the exit 0 at the end of the code as it would mask any other exit status of the script if it exited the loop.

---

Using the MIME-type of a file to select the correct grep tool:

for filename in "$HOST-mail-2018$1"[0-9][0-9]*; do
    case $( file -b -i "$filename" ) in
        text/plain*)          g=grep   ;;
        application/x-gzip*)  g=zgrep  ;;
        application/x-bzip2*) g=bzgrep ;;
        application/x-xz*)    g=xzgrep ;;
        *) printf 'Unknown filetype for "%s"\n' "$filename" >&2
           exit 1   # or continue or break
    esac

    "$g" 'Cannot .*TLS' "$filename"
done

This would correctly pick the correct grep tool regardless of what the filename suffix was (as long as the file was of one of the supported filetypes). I've also added plain grep for ordinary text files.

Related:

• Why does my shell script choke on whitespace or other special characters?
• Why *not* parse `ls` (and what do to instead)?