extract field since awk is very awkward for me.

extract-fields.rb

#!/usr/bin/env ruby
require 'optparse'
require "pp"

class CLI
  def parse_options(argv = ARGV)
    op = OptionParser.new
    
    self.class.module_eval do
      define_method(:usage) do |msg = nil|
        puts op.to_s
        puts "error: #{msg}" if msg
        exit 1
      end
    end
    
    # default value
    opts = {
      report: false,
      split: "\t",
      join: "\t",
      fields: [],
    }
    
    op.on('-r', '--report', "Report field configuration from very 1st line. (default: #{opts[:report]})") {|v| opts[:report] = v }
    op.on('-s', '--split VALUE', "string value (default: #{opts[:split].inspect})") {|v|
      opts[:split] = v.gsub('\\t', "\t")
    }
    op.on('-j', '--join VALUE', "string value (default: #{opts[:join].inspect})") {|v| 
      opts[:join] = v.gsub('\\t', "\t")
    }
    op.on('-f', '--fields one,two,three', Array, "fields to extract (default: #{opts[:fields]})") {|v| opts[:fields] = v }
    
    begin
      args = op.parse!(argv)
    rescue OptionParser::InvalidOption => e
      usage e.message
    end
    
    [opts, args]
  end
  
  def run
    opts, args = parse_options
    
    if (opts[:report])
      puts "opts: #{opts.to_s}"
      puts "args: #{args.to_s}"
      puts
      ARGF.each do |line|
        line.chomp.split(opts[:split]).each_with_index do |e, idx|
          puts "%2d: #{e}" % (idx + 1)
        end
        break
      end
      puts
      return
    end
    
    fields_to_extract = opts[:fields].map {|n| n.to_i }
    ARGF.each do |line|
      split = line.chomp.split(opts[:split])
      extracted = fields_to_extract.map {|n| split.values_at(n-1) }
      puts extracted.join(opts[:join])
    end
  end
end

CLI.new.run

USAGE.md

help

$ ruby extract-fields.rb -h
Usage: extract-fields [options]
    -r, --report                     Report field configuration from very 1st line. (default: false)
    -s, --split VALUE                string value (default: "\t")
    -j, --join VALUE                 string value (default: "\t")
    -f, --fields one,two,three       fields to extract (default: [])

Now explain with this sample.txt each fields are separated with tab(\t) char.

$ cat sample.txt
foo1	foo2	foo3
bar1	bar2	bar3
baz1	baz2	baz3

-r is useful to check field number from very 1st line.

$ cat sample.txt | ruby extract-fields.rb -r
opts: {:report=>true, :split=>"\t", :join=>"\t", :fields=>[]}
args: []

 1: foo1
 2: foo2
 3: foo3
 

Now lets' extract field 1 and 3.

$ cat sample.txt | ruby extract-fields.rb -f 1,3
foo1	foo3
bar1	bar3
baz1	baz3

This time, I extract 1 and 3, but custom order.

$ cat sample.txt | ruby extract-fields.rb -f 3,1
foo3	foo1
bar3	bar1
baz3	baz1

With -j option, I can join fields with custom string, here I use --.

$ cat sample.txt | ruby extract-fields.rb -f 3,1 -j '--'
foo3--foo1
bar3--bar1
baz3--baz1
$