Parse all column elements from a linux bash command output

Question

I would like to parse the column elements of the output from the command lsscsi.

Here is a sample output,

# lsscsi

[0:0:0:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sda
[0:0:1:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sdb
[1:0:1:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sdc

Example if I want column 2, my output should be,

disk
disk
disk

If cloumn 7,

/dev/sda
/dev/sdb
/dev/sdc

Thanks

anubhava · Accepted Answer · 2022-08-31T14:37:22.220

11

Use awk like this:

awk -v col=7 '{print $col}' file

Or to print 2 columns:

awk -v col1=2 -v col2=7 '{print $col1, $col2}' file

OR to make it print multiple columns using a colon delimited list:

awk -v col='2:7' '
BEGIN {n = split (col, arr, /:/)}
n {
   for (i=1; i in arr; ++i)
      printf "%s%s", $arr[i], (i < n ? OFS : ORS)
}' file

edited Aug 31 '22 at 14:37

answered Dec 04 '13 at 17:47

anubhava

761,203
64
569
643

Thanks that works! out of curiosity, so what if i need both columns 2 & 7? – user2887201 Dec 04 '13 at 17:50
1

Why not `awk '{print $7}'`? – Eric Wolf Jul 22 '18 at 16:20
Sure, that can be done too. This is just passing column no from command line – anubhava Jul 22 '18 at 16:29

score 5 · Answer 2 · answered Dec 04 '13 at 18:33

5

With bash you can use 'read -a', this reads a line from standard input, splits using IFS field separator which has space by default and populates array variable which can be referenced as ${array[n]} (zero based).

while read -a arr; do
    echo "${arr[1]}"
    echo "${arr[6]}"
done

answered Dec 04 '13 at 18:33

Mindaugas Kubilius

343
1
8

what is `arr` in this case? – sf8193 Mar 30 '22 at 22:39
@sf8193 arr is an array variable which will hold input from stdin splitted by IFS – Mindaugas Kubilius Apr 14 '22 at 18:40

score 2 · Answer 3 · answered Oct 19 '17 at 18:40

Unfortunately lsscsi does not provide consistent parseable output nor the chance to change the field separator as far as I see. With two different harddisks on the same system the column for the device is differing by one! The following code sniplet should give you an idea about what I mean and how I worked around it... at least to get the device which is at the end of the line. To get the product name and the product revision correctly, it could also be an idea to call each device with lssci -c a:b:c:d and then parse this output.

lsscsi > /tmp/tmp$$.out
while read line
do
    echo $line
    echo 012345678901234567890123456789012345678901234567890123456789
    echo 0.........1.........2.........3.........4.........5.........
    ID=$(echo $line | cut -c2)
    TYPE=$(echo $line | cut -c16-18)
    PRODUCT=$(echo $line | sed -n 's/.\{18\}\(.*\) *\/dev\/.*/\1/p')
    DEVICE=$(echo $line | sed -n 's/.\{18\}.*\/dev\/\(.*\)/\1/p')
    echo $ID-$TYPE-$DEVICE-$PRODUCT
done

would be easier to read if you use a different separator: `'s#.\{18\}$.*$ */dev/.*#\1#p'` — phuclv, Dec 21 '19 at 05:21

denten · Answer 4 · 2015-12-03T04:03:09.293

1

Another easy way to do it without awk:

lsscsi | tr -s ' ' | cut -d' ' -f7

The last flag specifies the column.

edited Dec 03 '15 at 04:03

answered Dec 02 '15 at 16:23

denten

691
1
6
15

fails : lsscsi has a fixed column width and visibly not field separator. – MUY Belgium Feb 10 '20 at 11:01
Tested it again on a Linux machine and it works fine. My lsscsi output only has four columns, and therefore I use `-f4` for the final argument. All output is treated as text in the shell, and in this case, we force separation by single space. – denten Apr 27 '20 at 17:46
this worked for me getting the 3rd column of docker image table in terminal – misinglink Jun 10 '21 at 06:01
it fails because the name field may contain spaces, and the path can also contain spaces although rarer – phuclv May 16 '23 at 09:29

score 1 · Answer 5 · answered May 03 '21 at 03:38

Simple examples:
(Each example is independent.)

Select the column 2:

echo -e "[0:0:0:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sda \
  \n[0:0:1:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sdb \
  \n[1:0:1:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sdc" \
  | awk '{print $2}'

Select the column 7:

echo -e "[0:0:0:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sda \
  \n[0:0:1:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sdb \
  \n[1:0:1:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sdc" \
  | awk '{print $7}'

Select the column 2 & 7:

echo -e "[0:0:0:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sda \
  \n[0:0:1:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sdb \
  \n[1:0:1:0]   disk   ATA   VBOX HARDDISK   1.0   /dev/sdc" \
  | awk '{print $2, $7}'

Results:

score 0 · Answer 6 · answered Feb 10 '20 at 11:31

You may want to use the separation in columns with lsscsi : Use perl to insert tab and trim the fields.

lsscsi -g \
| perl -ne 'foreach$c(qw (64 53 47 30 21 13)){
              substr$_,$c,0,"\t"
            };
            s/ *([\t\n])/\1/g;  # trim the fields
            print'

Or one liner :

 lsscsi -g|perl -ne 'foreach$c(qw(64 53 47 30 21 13)){substr$_,$c,0,"\t"};s/ *([\t\n])/\1/g;print'

Parse all column elements from a linux bash command output

6 Answers6

Select the column 2:

Select the column 7:

Select the column 2 & 7:

Results: