tags:

views:

1777

answers:

2

Suppose I have a text file with data separated by whitespace into columns. I want to write a little shell script which takes as input a filename and a number N and prints out only that column. With awk I can do the following:

awk < /tmp/in '{print $2}' > /tmp/out

This code prints out the second column.

But how would one wrap that in a shell script so that a arbitrary column could be passed in argv?

+5  A: 
awk -v x=2 '{print $x}'
or in a shell script:
#!/bin/sh
num=$1
awk < /tmp/in -v x=$num '{print $x}' > /tmp/out
Ray
Thanks for the edit Jonathan...
Ray
+1  A: 
awk '{print $'$myvar'}' < /tmp/in > /tmp/out

Where $myvar is your variable column (an integer). Watch out for script injections!

strager
Or: awk "{print \$$myvar}" ... the advantage being fewer quotes.
Jonathan Leffler