AWK output to bash Array

Question

Im trying to put the contents of a simple command in to a bash array however im having a bit of trouble.

df -h | awk '{ print  $5" "$6 }'

gives percentage used in the file systems on my system output looks like this:

1% /dev
1% /dev/shm
1% /var/run
0% /var/lock
22% /boot
22% /home
22% /home/steve

I would then like to put each of these lines into a bash array array=$(df -h| awk '{ print $5 $6 }')

However when I print out the array I get the following:

5%
/
1%
/dev
1%
/dev/shm
1%
/var/run
0%
/var/lock
22%
/boot
22%
/home
22%
/home/steve

Bash is forming the array based on white spaces and not line breaks how can i fix this?

What do you want to do with the array after having it in bash ? I mean ... probably you can do that in awk ... — Dr. belisarius, Oct 29 '10 at 22:17
Well the next part of the script would be to run something similar to this command "find . -type f -exec du -sh {} \; | sort -h" on the chosen file system (user would choose a file system that is filling up too much from the array in part one). I would like to modify this second command to just show the 10 biggest files in the file system so the user can choose if to delete them or not — , Oct 29 '10 at 22:26

score 10 · Accepted Answer · answered Oct 29 '10 at 22:17

10

You need to reset the IFS variable (the delimeter used for arrays).

OIFS=$IFS #save original
IFS=','
df -h | awk '{ print $5" "$6"," }'

answered Oct 29 '10 at 22:17

Brian Clements

1

You're welcome. If this was the right answer for you, you should upvote and mark as the accepted answer. – Brian Clements Oct 29 '10 at 23:34
3

You can omit the comma in the AWK command and set `IFS` to newline: `IFS=$'\n'` (don't forget `IFS=$OIFS` afterward). – Dennis Williamson Oct 29 '10 at 23:41
1

I've noticed that I sometimes get odd behavior when I use newline for ifs. I've gotten in the habit of using comma. – Brian Clements Oct 30 '10 at 00:38

score 1 · Answer 2 · answered Oct 30 '10 at 06:48

1

You could do this in Bash without awk.

array=()
while read -a line; do
    array+=("${line[4]} ${line[5]}")
done < <(df -h)

answered Oct 30 '10 at 06:48

ephemient

score 0 · Answer 3 · answered Oct 29 '10 at 22:36

0

You may use this:

eval array=( $(df -h | awk '{ printf("\"%s %s\" ", $5, $6) }') )

answered Oct 29 '10 at 22:36

1

What's the `eval` for? Just leave off the extra quotes in the AWK command and do the array assignment directly. – Dennis Williamson Oct 29 '10 at 23:34

3 Answers3