awk using dynamic column number

Question

I would like to have a bash function that prints the specific column of a given string.

My solution works fine but it is verbose and looks ugly to me.

Code:

#!/bin/bash

function get_ip_segment() {
    case "$2" in
        1)
            printf "%s" "$(echo $1 | awk -F '.' '{print $1}')"
            ;;
        2)
            printf "%s" "$(echo $1 | awk -F '.' '{print $2}')"
            ;;
        3)
            printf "%s" "$(echo $1 | awk -F '.' '{print $3}')"
            ;;
        4)
            printf "%s" "$(echo $1 | awk -F '.' '{print $4}')"
            ;;
        *)
            printf "%s" "$1"
            ;;
    esac
}

IP="12.34.56.78"
echo $IP

echo "$(get_ip_segment $IP 1)"
echo "$(get_ip_segment $IP 2)"
echo "$(get_ip_segment $IP 3)"
echo "$(get_ip_segment $IP 4)"

Output:

12.34.56.78
12
34
56
78

Then I tried to optimize this code by removing the switch structure, but it printed an unexpected output.

printf "%s" "$(echo $1 | awk -F '.' '{print column_id}' column_id=$2)"

Output:

But somehow, the variable substitution does not work correctly inside awk. That could be great if I could have a one-line long function instead of this verbose one.

anubhava · Accepted Answer · 2023-04-28T11:18:18.310

2

You can surely do it like this:

get_ip_segment() {
   awk -F. -v c="$2" '{print 1<=c && c<=NF ? $c : $0}' <<<  "$1"
}

IP="12.34.56.78"

get_ip_segment "$IP" 1
12

get_ip_segment "$IP" 2
34

get_ip_segment "$IP" 3
56

get_ip_segment "$IP" 4
78

get_ip_segment "$IP"
12.34.56.78

get_ip_segment "$IP" 5
12.34.56.78

edited Apr 28 '23 at 11:18

answered Apr 28 '23 at 10:19

anubhava

761,203
64
569
643

this condition`1<=c`, you can just check it once at `BEGIN { }`, and if it failed, set it to some ridiculously large number like `8^16` so it would always fail `c<=NF` (but this extra stuff is only worth it if the input might be more than a handful of rows) – RARE Kpop Manifesto May 02 '23 at 10:27
Yes, nice optimizations ! – anubhava May 02 '23 at 10:31
1

thanks. if you wanna streamline it further, perhaps :::::::: ::::::::::::::::: :::::::::: :::::::::::::::: ::::::::: ::::: :::::::::::::::: `{ print $((c <= NF) * c) }'` ::::::::: replacing the branching ternary with a multiplication – RARE Kpop Manifesto May 02 '23 at 11:36
That's really cool ! – anubhava May 02 '23 at 13:46
keep the sarcasm – RARE Kpop Manifesto May 02 '23 at 15:41

M. Nejat Aydin · Answer 2 · 2023-04-28T13:44:24.853

In plain Bash:

get_ip_segment() {
    local output=$1
    if [[ $2 = [1-4] ]]; then
        local -a octets
        IFS=. read -ra octets <<< "$output"
        output=${octets[$2 - 1]}
    fi
    printf '%s\n' "$output"
}

ip=12.34.56.78
get_ip_segment "$ip" 1
get_ip_segment "$ip" 2
get_ip_segment "$ip" 3
get_ip_segment "$ip" 4
get_ip_segment "$ip"

Or, in POSIX shell:

get_ip_segment() {
    case $2 in
        1) output=${1%%.*};;
        2) output=${1#*.}
           output=${output%%.*};;
        3) output=${1%.*}
           output=${output##*.};;
        4) output=${1##*.};;
        *) output=$1
    esac
    printf '%s\n' "$output"
}

ip=12.34.56.78
get_ip_segment "$ip" 1
get_ip_segment "$ip" 2
get_ip_segment "$ip" 3
get_ip_segment "$ip" 4
get_ip_segment "$ip"

score 1 · Answer 3 · answered Apr 28 '23 at 10:57

Using any awk in any POSIX shell:

$ cat tst.sh
#!/usr/bin/env bash

get_ip_segment() { awk -v ip="$1" -v seg="$2" 'BEGIN{ split(ip,a,"."); print (seg in a ? a[seg] : ip) }'; }

ip=12.34.56.78
get_ip_segment "$ip" 1
get_ip_segment "$ip" 2
get_ip_segment "$ip" 3
get_ip_segment "$ip" 4
get_ip_segment "$ip" 0
get_ip_segment "$ip" 5

$ ./tst.sh
12
34
56
78
12.34.56.78
12.34.56.78

score 0 · Answer 4 · answered May 02 '23 at 05:19

0

$ ip=12.34.56.78
$ get_ip_segment(){ cut -d. -f $2 <<<$1; }
$ for i in {1..4}; do get_ip_segment $ip $i; done
12
34
56
78

answered May 02 '23 at 05:19

ufopilot

3,269
2
10
12

awk using dynamic column number

4 Answers4