bash - Shuffle piped through sed gives different number of lines -

i have data file thousands of lines, each consisting of 5 numbers. example:

23 31 56 21 34 34 76 34 75 32 ... ... 

i want write bash script select n% lines @ random, , output them last entry set 0. remainder of entries want output line is. don't care in order lines output.

my attempt @ doing shuffle file, take first n% of lines , use awk print them 0 in last place. output remainder of lines. here attempt:

#! /bin/bash number=$2 numlines=$(less $1 | wc -l) number=$(echo $number'*'$numlines | bc) number=$(echo $number'/'100 | bc)  shuffledfile=$(less $1 | shuf) # following line echos shuffled file, gets first $number lines, , prints them 0 in final column echo "$shuffledfile" | sed -n --unbuffered "1,/$number/p" | awk '{print $1" "$2-7200" "$3" "$4" 0"}' echo "$shuffledfile" | sed -n "/${number}/,/${numlines}/p" | awk '{print $1" "$2" "$3" "$4" "$5}' 

my problem each time run script different number of lines output. have determined if don't shuffle file, works expected. in advance.

you using wrong notation printing lines sed, should be:

sed -n 'fromline,toline p' 

currently printing line 1 whichever line contains /$number/, or in second case first line containing /${number}/ following line containing /${numlines}/ which, random input, rather unpredictable.