Various Linux commands can present file and directory sizes in “human readable” format, where they append a “K” or “M”, etc, to indicate the file size; for example:
1 2 3 4 5 6 | # du -sh * 960K auth.log.1 124K daemon.log 8.0K dpkg.log.5.gz 2.1M mcollective.log.3 56K messages |
If we want to sort them by size, the sort command by itself isn’t much help. That’s because, by default, sort implements a lexical sort, and thus “2.1M” is sorted above “56K”, despite human beings recognising that the second is smaller:
1 2 3 4 5 6 | # du -sh *|sort 124K daemon.log 2.1M mcollective.log.3 56K messages 8.0K dpkg.log.5.gz 960K auth.log.1 |
In a short list, it’s easy to see that the mcollective.log.3 file is the largest, but what if we have a long list? The sort command now takes a “-h” qualifier which understands the “human readable” format:
1 2 3 4 5 6 | # du -sh *|sort -h 8.0K dpkg.log.5.gz 56K messages 124K daemon.log 960K auth.log.1 2.1M mcollective.log.3 |
Sorted, in more ways than one.
Could this Linux Tip be improved?
Let us know in the comments below.
