awk Command Loop

The way awk works is that by default it steps through each line of text and splits every space-separated item into fields, in this case we have fields $2, $3,and $4 of interest. Now what if we could store each $2 field of each line into a list and then print it ? That's where arrays can help.

Treating the body of an awk code as one big while loop, we can separate fields 2,3,and 4 into appropriate arrays. With each line, arrays will get filled. So we basically sort everything into columns(arrays) as we go though lines. Once done, we can go through a loop printing each colum(array) separately, while just adding text "Column#".

$ awk '{ array1[NR]=$2;array2[NR]=$3;array3[NR]=$4} END{for(i=1;i<=NR;i++){print "Column"i" "array1[i]};printf "\n";for(i=1;i<=NR;i++){print "Column"i" "array2[i]}; printf"\n"; for(i=1;i<=NR;i++){print "Column"i" "array3[i] };printf "\n" }' columns.txt           
Column1 a
Column2 1
Column3 x

Column1 b
Column2 2
Column3 y

Column1 c
Column2 3
Column3 z

Of course that one long line is a bit awkward to use (no awk puns intended). We can put everything into a script:

#!/usr/bin/awk -f

{ array1[NR]=$2;array2[NR]=$3;array3[NR]=$4 } 

END{
  for(i=1;i<=NR;i++){
     print "Column"i" "array1[i]};printf "\n";
   for(i=1;i<=NR;i++){
     print "Column"i" "array2[i]}; printf"\n"; 
   for(i=1;i<=NR;i++){
     print "Column"i" "array3[i] };printf "\n" 
}

Call that script columnate.awk, change permissions to executable with chmod +x columnate.awk and run it with any text file as a parameter:

$ ./columnate.awk columns.txt                                                                                                                          
Column1 a
Column2 1
Column3 x

Column1 b
Column2 2
Column3 y

Column1 c
Column2 3
Column3 z