I have a file, which looks like:
Lorem ipsum dolor sit amet... 2465m id porttitor libero mauris at magna... 1m istique pretium tincidunt. V...1200m ... I want to calculate (using unix commands, awk, grep etc.) sum of trailing number, I mean (2465 + 1 + 1200 + ...) / NUMBER_OF_LINES. What is the best way to do that?
This makes it:
$ awk '{sum+=$NF+0} END{print "total sum is " sum " and average " sum/NR}' file total sum is 3666 and average 1222 $NF stands for the last word of every line. It uses $NF+0 so that the trailing m and other characters are not taken into account. Then it keeps the sum in sum variable, that is printed at the end. Finally it prints the average, taken dividing the sum by NR, which has the value of number of lines processed.
... would be filled with other content.I would do this:
grep -oP '\d+(?=m\s*$)' file|awk '{_+=$0}END{printf "Avg: %0.2f\n",_/NR}' it has two processes, grep and awk, but it works the case that the number column/field starts with non-digit char. like:
foo123 456 ffffff100m xbar 222 444 bbbbb200m 234 df343 xxxxxx300m with the above example, it outputs :
Avg: 200.00 with your example it outputs:
Avg: 1222.00 ... is confusing. if there were spaces, yours and konsoleboxs answer are better.awk '{sum+=$NF}END{printf "Average: %0.2f\n",sum/NR}' file Input:
Lorem ipsum dolor sit amet... 2465m id porttitor libero mauris at magna... 1m istique pretium tincidunt. V...1200m Output:
Average: 822.00 Some parts actually didn't have spaces and does not work with FS=" ." so you really had to convert the string.
awk '{t=$NF;gsub(/[^0-9]/,"",t);sum+=t}END{printf "Average: %0.2f\n",sum/NR}' Output:
Average: 1222.00 
echo "(2465+1+1200)/3"|bc gives 1222 on my laptop. the big "V" in last line is the problem. your awk works only if the NF beginning with the number.foo300bar300m ?
foo...bar... foo200m? I mean non-blank char before the first digit?