I need to manipulate an output file C2K248MX - 3 C2K248MX - 3 C2K248MX - 3 C2K248NE - 1 C2K248NN - 2 C2K248NN - 2 C2K248OH - 2 C2K248OH - 2 C2K248OM - 1 C2K248PB - 1 C2K248PD - 1 I only want to see 1 entry with a count rather than an entry for each count,would also like to see the highest first Any help would be appreciated as Im not a Unix professional and only have limited knowledge... Many thanks

A Korn shell, break point report is one way of doing it. I'm assuming the file is already sorted and is 3 fixed length fields: #!/bin/ksh # no error checking rm -f new.file buffer="" sum=0 while read f1 f2 f3 do if [[ -z $buffer ]] then buffer="$f1" sum=$f3 continue fi if [[ "$f1" != "$buffer" ]] then echo "$buffer $sum" sum=$f3 else ((sum=sum+$f3)) fi buffer="$f1" done < out.file > new.file echo "$buffer $sum" >> new.file # end script by default, the sort command sorts in descending order: sort new.file

Many thanks for the prompt reply Not sure if I can combine this with current script in use. Current script: #!/usr/bin/ksh EVTLOG=evtlog.`date +"%Y-%m-%d"` (Sets date back to previous day) if [ $# -ne 1 ] then # Assume using todays log logfile=$LOGS/mq/$EVTLOG else logfile=$1 fi flipped=`cat $logfile | grep -i -c "Dual Site Indicator changed"` print "Operating with file : $logfile" print "Total Site Switches : $flipped " for atmflip in $(cat $logfile | grep -i "Dual Site Indicator changed" | cut -c382-389 | sort) do atmflipcount=`cat $logfile | grep -i "Dual Site Indicator changed" | grep -c $atmflip` print "$atmflip - $atmflipcount" done. That last print statement produces the results you see at the begining need to arrange that to produce a list with entry and count only - Many thanks for trying though

Myself, I'm very careful with changing production code so I try to limit changes when I can. You don't say how the output of your Current script is captured - a file perhaps? If so, you could use that file as the input file of the script I provided. I called the file out.file. Another way is to encorporate my changes into your current script. To minimize these changes, send the output of your last for loop to a temporary file: for atmflip in $(cat $logfile | grep -i "Dual Site Indicator changed" | cut -c382-389 | sort) do atmflipcount=`cat $logfile | grep -i "Dual Site Indicator changed" | grep -c $atmflip` print "$atmflip - $atmflipcount" done > /tmp/out.file # note sending output to /tmp/out.file Now, add this code to the end of you Current script: # untested rm -f /tmp/new.file buffer="" sum=0 while read f1 f2 f3 do if [[ -z $buffer ]] then buffer="$f1" sum=$f3 continue fi if [[ "$f1" != "$buffer" ]] then echo "$buffer $sum" sum=$f3 else ((sum=sum+$f3)) fi buffer="$f1" done < /tmp/out.file > /tmp/new.file echo "$buffer $sum" >> /tmp/new.file # end script If the output of your Current script is presently captured in a file then you can eliminate redirecting to /tmp/new.file

Hi Robert, Sorry for the late reply, Not sure if my suggestion will be of any help. Here is a one-liner for what u have asked for. Your file is:- C2K248MX - 3 C2K248MX - 3 C2K248MX - 3 C2K248NE - 1 C2K248NN - 2 C2K248NN - 2 C2K248OH - 2 C2K248OH - 2 C2K248OM - 1 C2K248PB - 1 C2K248PD - 1 Now here is the oneline solution for your problem:- bash-2.03# cut -d"-" -f1 list_2|sort|uniq -c|sort -nr 3 C2K248MX 2 C2K248OH 2 C2K248NN 1 C2K248PD 1 C2K248PB 1 C2K248OM 1 C2K248NE Hope its of any use to u, Regards, Sujan Banerjee