如何在shell中使用awk变量

Question

I am having the following structure(only one row)

col1  col2  col3  col4
----------------------
12    34    14    10

what I have to do is I have to fetch col2 and col4 and used that value in shell.

a=$(awk 'NR<2{$2}')
b=$(awk'NR<2{$4}')

and I have to use a and b in my shell scripting. but this is not a good idea as I am using awk two times instead I am fetching value like this .

awk'NR<2{ a=$2;b=$4}'

but I dont know how can I use this a and b in my shell.

Answer 1

with eval you could assign more shell variables in one awk command. see the test below:

kent$  echo $a,$b
,

kent$  cat test
col1  col2  col3  col4
----------------------
12    34    14    10

kent$  eval $(awk 'NR==3{print "a="$2;print "b="$4}' test)

kent$  echo $a,$b
34,10

note that I just copied your input example, with header/titles etc. so in my script the data line is NR=3. you can adjust it if it is not in your case.

Answer 2

The read built-in is useful here.

read _ a _ b _ < <( tail -n +3 data.txt)  # Assign the columns you don't care about to _
echo $a,$b

This avoids the use of eval , which should generally be avoided if you can find another option.

Answer 3

$ v=( $(awk 'NR==3{print $2 ORS $4}' file) )
$ a="${v[0]}"
$ b="${v[1]}"
$ echo "$a,$b"
34,10