Shell Programming and Scripting

I guess I still don't understand something.....
You're trying to calculate the average per column. The column average is: sum of the cells DIVIDED by the number of rows.
In your example you have ONE column and 4 rows:

Code:

1
2
1
2

the sum is: 6
# of rows: 4
------------
avg = (6/4) = 1.5

I must be missing something obvious......

Sorry maybe I'm not made my self clear. So I will explain.

I've a list in a file that looks like this:

1
2
1
2

What I want to do is to tell my script how many columns I would like. So if I want, let's say, 2 columns it would look like this:
1 2
1 2

And then I would like to add each element in a column and divide it by the number of elements in that column and then print out the average. So my output should be in the case above:

1 2

Where 1 is the average value of column 1 and 2 is the average value of column 2.

And I should add that I'm using Solaris 9 and awk is the only command available to me.

ok, NOW I think I understand:

# this will split it into '2' columns and do the avg - '2' is the default # of columns
nawk -f bag.awk myList

# this will split it into '3' columns and do the avg
nawk -v n=3 -f bag.awk myList
bag.awk:

Code:

BEGIN {
    n=(n=="") ? 2 : n
}
{
   idx=(FNR%n) ? FNR%n  : n
   sum[idx] += $0
   nf=(idx>nf) ? idx : nf
   nr=((FNR%n) == 1) ? nr+1 : nr
}

END {
    for(i=1; i<=nf; i++)
        printf("%s%s", sum[i]/nr, (i==nf) ? "\n" : OFS)
}

Works like a charm, thank you. But if the input should come from another command, how would that look like?

output | awk -v n=2 -f bag.awk

Would the above work? Because the input isn't always a file, sometimes it is a computation from a previous command. But your script was the first I got to work without any fuss.

you got it - that's the syntax.

I wonder why I have to use nawk when sitting in Solaris 9, since it seems to have both awk and nawk?

>>which awk
/usr/bin/awk

>>which nawk
/usr/bin/nawk


Because at home I used awk for the script vgersh99 gave me sitting on Mac OS X (latest version 10.4.xx), but when using a friends Solaris 9, I have to use nawk.

What is the difference? And why does it seem to work with awk on Mac OS X and only with nawk on Solaris ?

Just interested to know the difference for future knowledge.

Solaris' '/usr/bin/awk' is an acient broken implementation of 'awk'. The prefered awk version on solaris is either '/usr/bin/nawk' or a POSIX-complient implementation under '/usr/xpg4/bin/awk' (or 'gawk' if you have one).
Here's the AWK Compatibility List.