[ILUG] ugly df

[Brian Foster]

  | Date: Sat, 11 Nov 2006 07:51:20 +0900
  | From: "Josh Glover" <jmglov at wmalumni.com>
  | 
  | On 11/11/06, Paul Reilly <paulr at maths.tcd.ie> wrote:
  |  > Anyone know a simple way to make df display it's output nicely?
  | 
  | Sure. This fact that I call df twice is slightly garbage, but it gets
  | the job done and I cannot think of another way to not sort the first
  | line, short of using a tmp file.
  | 
  |   df -h | head -1; df -h | sed -e 1d | sort

 sorry, I couldn't resist the challenge to run df(1)
 just once ....

(df -h | sed -e '1{w/dev/fd/3
d;}' | sort -nk 1.9 ) 3>&1
Filesystem            Size  Used Avail Use% Mounted on
/dev/hdb1            1012M   38M  923M   4% /boot
/dev/hda4              23G   20G  2.8G  88% /a/hda4
/dev/hdb5              21G  4.4G   16G  22% /opt
/dev/hdb6              10G  2.1G  8.0G  21% /var
/dev/hdb7             6.1G  171M  5.9G   3% /
/dev/hdb9              11G  202M  9.9G   2% /tmp
/dev/hdb10             25G   11G   14G  45% /usr
/dev/hdb11             41G   18G   23G  45% /home
/dev/hdb12             41G   41G  571M  99% /a
tmpfs                 126M   12K  126M   1% /dev/shm

 this is not too general, does not do everything Paul
 wants (IIRC), has a race condition, and does not deal
 with multiple drives too well (as you can see in the
 above output), but does run `df' only once.  ;-)
 the embedded newline seems to be mandatory. ;-(

 except for the race condition, Josh's double-`df' has
 similar, but not identical, limitations.

 switching the tools used slightly (awk(1) is, in this
 case, possibly a better choice than sed(1)?) can be
 used to remove most(/all?) of these limitations.

cheers!
	-blf-
-- 
Experienced (20+ yrs) kernel/software Eng: | Brian Foster   Montpellier,
 • Unix, embedded, &tc;  • Linux;  • doc;  | blf at utvinternet.ie   FRANCE
 • IDL, automated testing, process, &tc.   |  Stop E$$o (ExxonMobile)!
Résumé (CV) http://www.blf.utvinternet.ie  |     http://www.stopesso.com