How to list directories size from a tar.gz archive using bash

Question

I have a huge tgz archive and I know there are several directories there, no files in the root.

I want to know exact size of directories files inside to estimate if they'll fit in my mounted volumes.

I've found this thread helping https://stackoverflow.com/a/11721660/1004388

Ilya Sheershoff · Answer 1 · 2018-10-05T13:30:33.403

This one-liner would do the trick:

tar tzvf /tmp/root.tgz | sed 's/ \+/ /g' | cut -f3,6- -d' ' | cut -f1 -d'/' | awk '{
    arr[$2]+=$1
   }
   END {
     for (key in arr) printf("%s\t%s\n", key, arr[key])
   }'

Example output:

usr 821233945
boot    11150620

Explanation:

tar tzvf filename - lists all files in archive in a ll -r style
sed contracts multiple spaces into single to help cutting
first cut cuts third and sixth fields and leaves everything after the sixth field intact, considering the delimiter is space - now we have size in first column and file path in second column
second cut: since we need only top-level entries, we cut off with the first directory separator - only first field, considering / is the separator
awk is used to group by second field, summing the first one

I had to remove the z (mine wasn't gzipped) and change the second cut to -f1,2 because mine all started with ./ -- Very useful +1 — Rick, Nov 09 '21 at 21:24

score 0 · Answer 2 · answered Sep 06 '22 at 11:09

(Posting as a response, because I cannot comment yet.)

The response of Ilya above is correct, but on my system (Mac OS), the output of tar tzvf /tmp/root.tgz had a different column layout, so I had to replace the first sed and cut with awk '{ print $5,$9 }' (i.e. keep only the 5th and 9th column), giving this new one-liner:

tar tzvf /tmp/root.tgz | awk '{ print $5,$9 }' | cut -f1 -d'/' | awk '{
    arr[$2]+=$1
   }
   END {
     for (key in arr) printf("%s\t%s\n", key, arr[key])
   }'

How to list directories size from a tar.gz archive using bash

2 Answers2