How do you extract a particular column or row from the output of a command?

Question

Example:

abcd@abcd-abcde:~$ xinput --list --short

Virtual core pointer

↳ SynPS/2 Synaptics TouchPad                id=11   [slave  pointer  (2)]

↳ Logitech USB RECEIVER                     id=12   [slave  pointer  (2)]

How would I extract a column from the output, say the second one? And, for example, store it in a variable?

Is it possible to do this?

terdon · Accepted Answer · 2016-05-06T12:00:50.763

That's exactly what awk was made for. Using the output of xinput on my system:

$ xinput --list --short
⎡ Virtual core pointer                      id=2    [master pointer  (3)]
⎜   ↳ Virtual core XTEST pointer                id=4    [slave  pointer  (2)]
⎜   ↳ SIGMACHIP USB Keyboard                    id=12   [slave  pointer  (2)]
⎜   ↳ Logitech M325                             id=14   [slave  pointer  (2)]
⎜   ↳ AlpsPS/2 ALPS DualPoint Stick             id=17   [slave  pointer  (2)]
⎜   ↳ AlpsPS/2 ALPS DualPoint TouchPad          id=16   [slave  pointer  (2)]
⎣ Virtual core keyboard                     id=3    [master keyboard (2)]
    ↳ Virtual core XTEST keyboard               id=5    [slave  keyboard (3)]
    ↳ Power Button                              id=6    [slave  keyboard (3)]
    ↳ Video Bus                                 id=7    [slave  keyboard (3)]
    ↳ Power Button                              id=8    [slave  keyboard (3)]
    ↳ SIGMACHIP USB Keyboard                    id=11   [slave  keyboard (3)]
    ↳ Dell WMI hotkeys                          id=18   [slave  keyboard (3)]
    ↳ AT Translated Set 2 keyboard              id=13   [slave  keyboard (3)]
    ↳ Sleep Button                              id=9    [slave  keyboard (3)]
    ↳ UVC Camera (046d:0819)                    id=10   [slave  keyboard (3)]

$ xinput --list --short | awk -F"\t" '{print $2}'
id=2
id=4
id=12
id=14
id=17
id=16
id=3
id=5
id=6
id=7
id=8
id=11
id=18
id=13
id=9
id=10

awk will automatically split each input line into fields on the value given by -F. By default, it splits on whitespace, but here we need a tab. The 1st field is $1, the second $2 etc.

Alternatively, you can use cut:

$ xinput --list --short | cut -f2
id=2
id=4
id=12
id=14
id=17
id=16
id=3
id=5
id=6
id=7
id=8
id=11
id=18
id=13
id=9
id=10

You can save the output in a shell variable just like you would for any other command:

$ ids=$(xinput --list --short | awk -F"\t" '{print $2}')
$ echo $ids
id=2 id=4 id=12 id=14 id=17 id=16 id=3 id=5 id=6 id=7 id=8 id=11 id=18 id=13 id=9 id=10
$ echo "$ids"
id=2
id=4
id=12
id=14
id=17
id=16
id=3
id=5
id=6
id=7
id=8
id=11
id=18
id=13
id=9
id=10

You could also save it as an array:

$ ids=( $(xinput --list --short | awk -F"\t" '{print $2}') )
$ echo ${ids[@]}
id=2 id=4 id=12 id=14 id=17 id=16 id=3 id=5 id=6 id=7 id=8 id=11 id=18 id=13 id=9 id=10
$ echo ${ids[2]}
id=12

What if I want to get the first column? Will the arrows be included? Since I'm not using Linux atm I can't test right now.. — TellMeWhy, May 06 '16 at 13:52
@DevRobot fields are defined by the field separator. Here, I'm using `\t`, tab. Anything up to the 1st tab will be considered the first field, so yes, the arrows will be included. — terdon, May 06 '16 at 14:16

score 2 · Answer 2 · edited Apr 13 '17 at 12:25

Knowing that the items are tab separated, we can use shell-only way with $IFS variable and while read ; do . . .done structure.

$ xinput | while IFS=$'\t' read ONE TWO THRE ; do                                        
> echo $TWO
> done
id=2
id=4
id=11
id=3
id=5
id=6
id=7
id=8
id=9
id=10
id=12

To store items in a variable, we can use arrays, and

$ xinput | { while IFS=$'\t' read ONE TWO THRE ; do                                     
> array+=( "$TWO" )
> done ;
> echo ${array[@]} 
> }
id=2 id=4 id=11 id=3 id=5 id=6 id=7 id=8 id=9 id=10 id=12

Notice me using command | { command1 ; command2 ; command3 } structure. You will quickly learn that pipes start subshells and variables in subshells don't affect the main shell , thus you have to come up with some way of preserving those variables.

As for extracting a row , well that's fairly easy with grep:

    $ xinput | grep 'Virtual core keyboard'
    ⎣ Virtual core keyboard                     id=3    [master keyboard (2)] # extract line with specific string

    ↳ Toshiba input device                      id=12   [slave  keyboard (3)]
$ xinput | sed -n '3p' # extract 3rd line                                                          
⎜   ↳ SynPS/2 Synaptics TouchPad                id=11   [slave  pointer  (2)]

@Kos that does make up for simpler declaration and corrected my answer, however, for whatever reason I cannot remember to put `$` in front, hence using `printf` — Sergiy Kolodyazhnyy, May 06 '16 at 14:05

the_velour_fog · Answer 3 · 2016-05-06T11:03:17.263

1

to extract the column you could try awk

% xinput --list --short | awk '{print $3}'
core
Virtual
SynPS/2
core
core
Button
Bus
...

to assign that output to a variable a. note the value of this variable will be a string containing newlines after each word - not the easiest to work with. Probably better to assign to an array.

% a=$(xinput --list --short | awk '{print $3}')
vel@ubuntu:/media/michael/3ae9487f-1652-400e-8136-fe63519941ed% echo $a
core
Virtual
SynPS/2
core
core
Button
Bus
Button
USB2.0
WMI
Translated

edited May 06 '16 at 11:03

answered May 06 '16 at 10:57

the_velour_fog

2,290
5
23
34

1

I think the OP means the second output column, the ID field. You can get that with `awk -F"\t" {print $2}`. – terdon May 06 '16 at 11:01
@terdon cool, yes that was only my first iteration, but yeah you put the output into an array which is better for OP too – the_velour_fog May 06 '16 at 11:07

How do you extract a particular column or row from the output of a command?

3 Answers3