Question

I have an issue with using AWK to simply remove a field from a stream, illustrated below:

  1 int blah (void)   2 {   3         if (foo) {   4                 printf ("blah\n");   5         }          6         return 0;   7 }   

I use the following code to remove the first field:

$ awk '{ $1=""; print }' example.out 

 int blah (void)  {  if (foo) {  printf ("blah\n");  }  return 0;  } 

Why is this the case? Is this because AWK removes all whitespace – can this be prevented?

Answer 1

http://awk.freeshell.org/RangeOfFields

Contains a description how to do it. It also links to http://student.northpark.edu/pemente/awk/awktail.txt which contains 3 solutions to the problem. As far as i know, if you assign to a field, then the output field separator is used to concatenate all fields together. So ' '+ suddendly is collapsed to one space. Take it with a grain of salt though, i’m no awk expert. For example, try assigning : to the variable OFS, and colons instead of spaces will result in between fields in the output:

echo a b c | awk 'BEGIN{ OFS = ':' } { $1=''; print }' $ :b:c 

If you use gawk, then you can use its gensub extension which i find pretty straight forward to use:

print gensub($1 '[\t ]*(.*)', '\\1', 1);