summing contents of a field - bash

I have a file that contains data in the following format:

Field 1 Field 2 nCalls duration

My task is to sum the nCalls and dur开发者_运维问答ation fields. The code I have written so far is as follows:

cat ${file} |\
while read n1 n2 nCalls nDuration
do
  #sumCalls=$((${sumCalls} + ${nCalls}))
  #sumDuration=$((${sumDuration} + ${nDuration}))
  sumCalls=`expr ${sumCalls} + ${nCalls}`
  sumDuration=`expr ${sumDuration} + ${nDuration}`
  echo "${sumCalls} ${sumDuration}"
  echo -n "${appName} ${sumCalls} ${sumDuration}" > temp.txt
done

tail -n1 temp.txt >> ${outFile}  

Both the statements above do not work. I get syntax errors.

My question is:

1. Where am I going wrong with the above code?

2. Is there a better way to do this than write into and out of files?

Thanks,

Sriram

---

This is what awk was built for:

cat ${file} | awk '{c = c + $3; d = d + $4} END {print "APPNAME", c, d}'

The following transcript shows it in action:

pax$ export appName=xyz

pax$ echo 'a b 1 2
...> a b 3 4
...> a b 5 6' | awk -vAPPNAME=${appName} '{c=c+$3;d=d+$4}END{print APPNAME,c,d}'
xyz 9 12

---

A Ruby(1.9+) example

$ ruby -ane 'BEGIN{c=d=0};c+=$F[2].to_i;d+=$F[3].to_i;END{puts "APPNAME, #{c},#{d}"}' file

---

As already said, right tool would be awk.

To the errors: The code is fine and runs (esp both expressions). Possible errors:

1. You dont initialize your sum up variables to 0 (if they have already value than it could yield wrong results.
2. If your file contains more than 4 columns, your read behaves different: Read assigns the last passed value always the rest of the line! I.e. your last variable gets with a 5 column entry 2 space seperated values assigned => and here expr can give a syntax error. Solution for this: Add to the line a 5 variable REST. With 4 columns it stays empty, with more it catches up the rest.