How to parse df line-by-line?

Question

I would like to parse the output from df so I do

arr=( $(df -hPT | awk '{print $1, $2, $3, $5, $6, $7}') )
for f in ${arr[@]}; do echo ${f[*]};sleep 1;done

but I get

Filesystem
Size
Avail
Use%
Mounted
devtmpfs
7.8G
7.8G
0%
/dev
tmpfs
7.8G
7.7G
3%

where I had hoped for

Filesystem Size Avail Use% Mounted
devtmpfs 7.8G 7.8G 0%
/dev tmpfs 7.8G 7.7G 3%

My plan was that $f would be an array for each line, so I can manipulate each line and printf the output aferwards.

Question

Any ideas how I can get an array of the awk output per df line?

Answer 1

Solution 1st: In single while loop:

df -hPT | while read first second third fourth fifth sixth seventh
do
   echo $first $second $third $fifth $sixth $seventh
   sleep 1
done

Solution 2nd: Following single awk may help you on same.

df -hPT | awk '{print $1, $2, $3, $5, $6, $7;system("sleep 1")}'

Answer 2

在读取时使用...循环

df -hPT | while read line ; do echo "$line" | awk '{print $1, $2, $3, $5, $6, $7}' ;sleep 1 ; done

Answer 3

You can use the bash builtin lines to store the lines of a file/stream into an array:

mapfile -t lines < <(df -hPT | awk '{print $1, $2, $3, $5, $6, $7}')
for line in "${lines[@]}"; do
    manipulate "$line"
done

Need to redirect from the process substitution so we don't introduce any subshells.

You've seen what happens when you fail to quote the variable in the for loop: always quote your variables (unless you know exactly why not).

Answer 4

If you set IFS to \\n like that :

IFS=$'\n' arr=( $(df -hPT | \
  awk '{print $1, $2, $3, $5, $6, $7}') )
for f in ${arr[@]}; do
  echo ${f[*]}
done
echo "nb elt in arr = ${#arr[@]}"