4

I'm using unix to read through an excel file, sort by one column, and then I want a percentage (lets say the top 10%) of the sorted lines exported to a new .xls file. I have the code below which works fine but I need to do a wc -l prior to this line to get the "126", and then have to type it in (which is a hassle since I want to do successive filtering). 

cat /Desktop/Results.xls | sort -n -k 3 | awk -F "[\t]" '{printf "%s\n",$0}' | head -$((126*1/10)) >./Results_0167_1.xls
Rubén
  • 34,714
  • 9
  • 70
  • 166
user3470496
  • 141
  • 7
  • 33

1 Answers1

2

You can modify the awk to head the file itself:

sort -n -k 3 /Desktop/Results.xls| awk '{a[i++]=$0}END{for (j=1;j<=i/10;j++){printf "%s\n",a[j]}}' >./Results_0167_1.xls

The awk stores and counts ( a[i++]=$0 ) the records piped by sort, after this process , in the END part it will print the recors till 10% is reached: j<=i/10

Juan Diego Godoy Robles
  • 14,447
  • 2
  • 38
  • 52