Shell Programming and Scripting

Thank you Rudi, just adding to his comments in case you need each time sets like from 1 to 40, 41 to 80 and so on then you could nullify the values of variables named x and y as follows, whenever there is line completely divided by 40.

Code:

awk '{x+=$2;y+=$2^2;} NR%40==0{print sqrt(y/NR-(x/NR)^2);x=y=""}' Input_file

Again as I haven't seen sample Input_file so this is based on the previous threads communications and provided information only, hope this helps.


Thanks,
R. Singh

Thank you Rudic and Ravinder Singh.

Ravi's updated one is what I have been looking for.

But I think to calculate the average for every 40lines (treating 1 to 40 as set1, 41 to 80 as set2 etc), this command is right?

Code:

awk '{sum+=$3} (NR%40)==0{print sum/40; sum=0;}'

"Ravi's updated one" in post#8 is NOT what you've been looking for. Replace y/NR and x/NR by x/40 and y/40, respectively.

Rudic - In Post8 or Post2 of Ravi?

In post#8. And, add an END section for the residual lines (c.f. post#5)

I made it like this -

Code:

awk '{x+=$3;y+=$3^2;} NR%40==0{print sqrt(y/40-(x/40)^2);x=y=""}END {print sqrt(y/(NR%40)-(x/(NR%40))^2)}'

Is this what you meant Rudic?

Thank you so much!

Yes.

And, Yes to your question for the average in post#9, provided you add END {print sum/(NR%40