Grand totals in awk

Shell Programming and Scripting
1

I have a one-liner script like this that gives a total of everything in various directories:

for i in *;
do (cd $i && cd statelist && echo $i && ls -la |awk 'NR>3 {SUM += $5}\
 END {  print "Total number of elements " SUM }');done

It works just great but at the end I want to print a grand total of all the numbers in column 5. I tried to tack on another SUM statement but it didn't work. Not sure how this code should work for grand total.

2

It doesn't work because you are running awk 937 times to count 937 directories, counting 937 separate "grand" totals.

Also: If your 'one liner' is three terminals wide, it probably shouldn't be a one-liner.

Also you can simplify it a lot. The 'echo' is pointless (and may actually be a bug). You can glob for 'statelist' instead of searching for it in every possible file and folder too. And you don't even need to cd -- or loop at all, for that matter -- to ls -la it.

ls -la */statelist | awk '!/^(total|d)/ && (NR>5) {SUM += $5} END {  print "Total number of elements " SUM }'
1 Like
3

you use find -type d to insure that ${i} will only find directories

1 Like
4

That's a good idea, find will be simpler than ls here, but we want only files, not dirs...

find */statelist -type f -printf "%s\n" | awk '{ SUM+=$1 } END { print "sum " SUM }'

Where %s, to find, means 'size in bytes'. So we don't need to do any filtering in awk at all.