Frequency Counts Using a Shell Script

This shell script can generate frequencies and it includes options for choosing the records per case and subsetting based on that (for a hierarchical data file). It depends on another shell script (printrec0), linked below.

#! /bin/sh
#. Generate frequencies given a file, columns, and optional rec # and rec/case
case $# in
 0)   echo " ";
      echo " usage:   freq filename column(s) [r=record rc=records per case]" ;
      echo " example: freq test.dat 8-13 r=2 rc=5" ;
      echo " ";
      exit 0;;

# *)   echo " ";
#      echo " error:  ** syntax error/incorrect number of arguments **" ;
#      echo " usage:   freq filename column(s) [r=record rc=records per case]" ;
#      echo " example: freq test.dat 8-13 r=2 rc=5" ;
#      echo " ";
#      exit 1;;

 2)   echo " ";
      echo "freq of $1, column(s) $2";
      echo " ";
      echo "   freq value";
      echo " ";
      cut -c$2 $1 | sort | uniq -c > /tmp/frtmp$$;
      cat /tmp/frtmp$$;
      awk 'BEGIN {print "-------------"}
                 {n += $1} 
             END { printf("   %d cases\n",n)}' /tmp/frtmp$$;
      rm -f /tmp/frtmp$$;;
     
 4)   echo " ";
      echo "freq of $1, column(s) $2, record $3, $4 records per case";
      echo " ";
      echo "   freq value";
      echo " ";
      awk -f printrec0 $3 $4 $1 > /tmp/frtmp2$$;
      cut -c$2 /tmp/frtmp2$$ | sort | uniq -c > /tmp/frtmp$$;
      cat /tmp/frtmp$$;     
      awk 'BEGIN {print "-------------"}
                 {n += $1} 
             END { printf("   %d cases\n",n)}' /tmp/frtmp$$;
      rm -f /tmp/frtmp$$;
      rm -f /tmp/frtmp2$$;;
   
esac
echo " "
exit 0

(Note: credit for this shell script goes to Nelson Martinez).

---

Back to Frequency Count Page

Last modified: Sat Sep 6 17:21:22 EDT 2025