First two fields to be separated by _ and rest of the fields as they are

To change the whitespace between the first two columns to an underscore, I suggest sed:

 sed -e 's/[\t ]\+/_/' 

And if you were to need to ignore the header line:

sed -e '/^#/! s/[\t ]\+/_/' 

or, for the more general case (header might start with any char; \t works only with gnu sed)

sed -E '1! s/[[:blank:]]+/_/' 

As to the question about your awk code, the first print, should likely be a printf so as not to have it print an ill timed newline.

Starting from your code, this should give you the desired output:

awk ' NR>1 { printf( $1"_"$2 ); for (i=3;i<NF;i++) printf("\t%s", $i); printf("\n") } NR==1 { print } ' input > output 

This appears to work:

awk '{ if(NR>1) { printf $1"_"$2; for(i=3;i<NF;i++) {printf "\t"$i } } print "" }' input 

Here's a small Python 3 script, which does the job. The underlying premise here is to read each line character by character use two variables - one which tracks whether the first-to-second column underscores have been written, and another - which tracks whether we're permitted to substitute space with underscore.

I've noticed from OP's input file format that the second column is all numeric values. Thus, we can start with allowing for spaces to be substituted with underscores, but once we've written underscores and encountered a numeric character ( both conditions being true), we can turn off the write_ok variable, and the other spaces will be printed out as usual.

#!/usr/bin/env python3 import sys import os def count_first_spaces(string): write_ok = True underscores_ok = False for char in string: if char == " " and write_ok: print("_",end="") underscores_ok = True continue if underscores_ok and char.isdigit(): write_ok = False print(char,end="") print("") # add newline def main(): if not os.path.isfile(sys.argv[1]): sys.exit(1) with open(sys.argv[1]) as fd: for line in fd: if line.startswith('#'): print(line.strip()) else: count_first_spaces(line.strip()) if __name__ == '__main__': main() 

And here's the test run:

$ ./add_underscore.py input.txt #CHROM POS REF ALT ../S101_sorted.bam ../S102_sorted.bam ../S105_sorted.bam ../S107_sorted.bam ../S113_sorted.bam ../S114_sorted.bam ../S115_sorted.bam ../S Aradu.A01_______296611 T C T T T T T T T T T T T T T T T/C T T/C T T T T Aradu.A01_______326689 T C T/C T T T T/C T T T T/C T/C T T T T T T T T/C T/C T T Aradu.A01_______615910 T G T T T T T T T T T T T T T T T T T T T T T Aradu.A01_______661394 T A T T T T T T/A T T T T T T T T T T T T T T T 

If you want that data to be saved to a different file, run it as ./add_underscore.py input.txt > output.txt