Question

First, if there is a better title for this question, I am all ears!

Is there a better way of doing this?

I have a data file input.txt in this format:

field1=value1
field2=value2
field3=value3
.
.
.
field1=value4
field2=value5
field3=value6

and so on … and would like to end up with:

field1,value1 value4 ... valueM
field2,value2 value5 ... valueN
field3,value3 value6 ... valueO

What I’ve tried:

1. Generate unique keys: cat input.txt |awk -F"=" '{print $2}' |sort -u > data_key

2. Loop over rows in data_key

#!/bin/bash

file=input.txt
keys=`cat data_key`

for value in $keys
do
  output=`cat $file |grep $value |awk -F"=" '{print $2}' |tr -s '[:space:]' '[ *]' `
  echo $value, $output
done

Answer 1

The below awk-based solution should work. This uses associative arrays keyed by field names. Values are concatenated as they are encountered.

awk -F'=' '{z[$1]=z[$1]" "$2} END{for(i in z){print(i","z[i])}}' file_name.txt