How can I make this AWK command case-insensitive?

user_12899453 0
Viewed 645

I am trying to print occurrences of a field in a .csv file only using an awk command. For example in a file test.csv like so:

layla;rebel;TAT
han_solo;rebel;TAT
darth_vader;empire;DKS
yoda;rebel;TAT

with the command:

cat test.csv | ./how_many_are_we.sh dks

I wish to have the following output:

1

Here is my code in how_many_are_we.sh (which is working but case-sensitive):

#! /bin/bash
awk -F ";" -v location=$1 'BEGIN {count=0;} { if ($3 == location) count+=1} END {print count}'

I tried adding IGNORECASE=1 in different places but I can't seem to find the right way to make it work.

Excuse my bad wording, and thank you for the help.

awk shell linux bash
2 Answers

You could change the case of entered value and for 3rd field to lower case and then compare their values to make sure however they are entered comparison shouldn't be affected.

#!/bin/bash
awk -F ";" -v location="$1" 'BEGIN {location=tolower(location);count=0;} { if (tolower($3) == location) count+=1} END {print count+0}' Input_file

OR As per Glenn sir's comments use shell trick to make it lowercase in variable itself.

#!/bin/bash
awk -v location="${1,,}" 'BEGIN{FS=";"} (tolower($3) == location){count+=1} END{print count+0}' Input_file

OR more awksh way change awk command to following(above is OP's command fix this is to make it awksh style)

awk -v location="$1" 'BEGIN{location=tolower(location);FS=";"} (tolower($3) == location){count+=1} END{print count+0}'

NOTE: For using IGNORECASE=1, you should mention it either in BEGIN section like BEGIN{IGNORECASE=1} OR like an awk variable -v IGNORECASE="1".

Also on a side note, OP's shebang is having spaces between #! and /bin/bash which shouldn't be the case so I had fixed that too here.

user_58665800

Normally, if you have a script that can work easily on lower-case input, you can convert the input stream with tr, so it can also manipulate mixed-case input.

$ command | tr '[:upper:]' '[:lower:]' | ./script

However, if you want to mask the tr, you can embed it into your script by default using the following redirection:

#!/usr/bin/env bash
exec 0< <( tr '[:upper:]' '[:lower:]' )
# the rest of your script comes here:
awk '...'

which now allows you to do something like:

$ command | ./script
user_83440600
Related