Script returning total used space for every sub directory

[milkoholicBear]

I am trying to write a script that return the total used space for all sub directories 1 level down.
What I mean is imagine the following structure:
/home/bob contains d1 (3mb) , d2 (1mb) and d3 (5 mb)
home/bob/d1 contains d11 (1mb), d12 (10mb)
home/bob/d2 does not have any subdirs
home/bob/d3 contains d31 (35mb)

When I run my script against home/bob, it should return this:
d1 16mb
d2 1mb
d3 40mb
total 57mb

note 16mb for d1 is found by adding al the files in d1 (3mb) + all files in d11 (1mb) + all files in d12 (1omb)


when run in home/bob/d1 it should return
d11 1mb
d12 12mb
total 13mb



I tried a combo of for and du but I don't seem to be able to get to anything edible.

Any ideas ?

Thx in advance.
MB

[devils casper]

df -h --max-depth=1 /home/bob

[smolloy]

df -h --max-depth=1 /home/bob

Shouldn't that be du instead of df?

du -h --max-depth=1 /home/bob

[devils casper]

ouch.......
you are right as usual.
thanx again smolloy.

[milkoholicBear]

Thanks for your replies.
Indeed, du got the job done, df did not accept the --max-depth.
BM

[devils casper]

Indeed, du got the job done, df did not accept the --max-depth.

df is for calculating file-system disk space usage and du is for estimate file space usage.
i confuses in df and du sometimes..
well ! i'll try to remember from now on..

[tpl]

Quite handy, thanks for the idea--it's in my /usr/local/bin. I think the
report is readable. You could use awk's "length" function to replace
the magic numbers.

#/bin/bash
#Dirs--says how big are files/directories
#
du --max-depth=1 | sed 's/\.\///g;$d' | sort -k 2 |
awk '{a+=$1; printf "%20s\t%8d\n",$2,$1} END {printf "%20s\t%8d\n", "TOTAL",a}'

[ghostdog74]

Quite handy, thanks for the idea--it's in my /usr/local/bin. I think the
report is readable. You could use awk's "length" function to replace
the magic numbers.

#/bin/bash
#Dirs--says how big are files/directories
#
du --max-depth=1 | sed 's/\.\///g;$d' | sort -k 2 |
awk '{a+=$1; printf "%20s\t%8d\n",$2,$1} END {printf "%20s\t%8d\n", "TOTAL",a}'

actually, the total will be displayed at the last line of du --max-depth output. so there's no need to do counting again.

[makefunnow]

Hi I want to know is there any way with bash script to get from and to id with grep command in maillog of postfix

[wje_lf]

You don't want to put this request here, precisely. You want to click the "create new thread" link. And speaking of precisely, the more details you give us, the better we can help you.