Hacker News Reader

Comment by stinos | Hacker News Reader

Would be great if somebody could post Python/Ruby/.. code here that achieves the same, just to compare. Even with all the C++11 additions it never felt like scripting to me, and I use it practically every day.

Small nitpcik to the author: please define variables when you're going to use them, not C-style all at the beginning of the function. It makes code easier to understand. It doesn't make the reader wonder 'hey what's this variable going to be used for' then having to crawl through all code underneath it. It creates code that's easier to refactor. Also, but more arguably for such a small sample: http://stackoverflow.com/questions/1452721/why-is-using-name...

View on HN

Haskell:

  import System.Environment
  import Data.List

  main = do
    infile:outfile:_ <- getArgs
    input <- readFile infile
    writeFile outfile . unlines . sort . lines $ input

It's a bit odd way of matching items of a list, I'd prefer to specify it without any possible trailing items:

    [infile, outfile] <- getArgs

Also, the last two lines may be squeezed into one (though less readable):

    writeFile outfile =<< (unlines . sort . lines) `fmap` readFile infile

On your first point; I was matching the behaviour of the C++ program in the blog post, which allows any number of command line arguments but only uses the first two.

On your second point, I actually thought about writing the whole thing as

  getArgs >>= \a -> readFile (a!!0) >>= writeFile (a!!1) . unlines . sort . lines

but decided that that's exactly the kind of thing that gets Haskell programmers a bad reputation.

Agreed, it does and you made the right choice. But in reality once you know Haskell - bind is a very common operation and reading this version is very natural.

Here's a Perl example. This isn't optimally compact, but this is more or less how I would write a script like this if I had to put it into production at $work (maybe with an extra check to make sure that exactly two command line arguments were provided).

  use strict;
  use warnings;

  open my $read,"<",$ARGV[0] or die $!;
  my @lines = <$read>;
  close $read;
  open my $write,">",$ARGV[1] or die $!;
  print $write $line foreach my $line(sort @lines);
  close $write;

If version >= 5.10 you can drop the 'or die' and for that matter neither of the close statements are necessary(in any recent version).

Perl5 v12 and up:

    use v5.12;    # enables strictures
    open my $in,  "<", $ARGV[0];
    open my $out, ">", $ARGV[1];
    print $out sort <$in>;

Perl6:

    sub MAIN( $in, $out ) {
        spurt( $out, open($in, chomp => False).lines.sort.join )
    }

[EDIT: fixed typo in first example]

Here's a couple of interesting Perl alternatives:

  my @files = map { IO::File->new($_->[0], $_->[1]) || die $! } 
              ([shift, 'r'], [shift, 'w']);
  $files[1]->print( sort $files[0]->getlines );
  $_->close for @files;

And...

  use autodie;
  open my $out, '>', $ARGV[1];
  print {$out} sort do {open my $in, '<', $ARGV[0]; <$in>};

Ruby:

  array = IO.readlines ARGV[0]
  array.sort!
  File.write ARGV[1], array.join("\n")

edit: get input and output file names from command line

Now that's the succinctness I'd expect from an actual scripting language

Except that's not what the C++ code does. The C++ reads and writes from files given as command line arguments.

Fixed

Thanks. I know enough ruby to be dangerous so didn't know if more needed to be done than that simple substitution.

The thing is in ruby you are already in main method so there's no need to declare main function as an entry point. The main reason that the c++ version the code is longer has some historical/performance related issues! In c++11 it could've been with smaller standard library at design but that could break the old codes! Although obviously, "Little code =! Better code". What you want to achieve is actually more important. Btw, C++ as a scripting language? at first you might think that way but truly that's a big lie :)

I'm not arguing with that ;)

BTW, I would never use Ruby for anything that needs performance, but for ease of use and readability it's really great.

If we want to golf it, it would probably just be

  lines = IO.readlines ARGV[0]
  File.write ARGV[1], lines.sort.join('\n')

certainly more readable than the C++ even if you're unfamiliar with ruby.

Well, if you really want to golf it, you can just write:

  File.write ARGV[1], IO.readlines(ARGV[0]).sort.join('\n')

;)

You're fired.

    IO.write(ARGV[1], ARGF.lines.sort.join)

I believe you still need to join on newline, the default is space. However you can save a couple characters by using `$*` instead of ARGV.

Also, trim out the space between the arguments and kill the parenthesis for the optimal golfing.

Wow, this is really sweet and esoteric (didn't know about ARGF) - but, reading the docs for ARGF, wouldn't it also try to ready from ARGV[1]?

Maybe this would fix it:

  IO.write(ARGV.pop, ARGF.lines.sort.join('\n'))

Yes, good catch. join('\n') is wrong though - IO#lines preserves newlines, the default join is the empty string, and you meant "\n" :)

Here's the shortest I've gotten after looking at some of the comments:

  IO.write$*.pop,ARGF.lines.sort.join

And yes, it does work even without the space between `write` and `$*`.

Also after testing, I realized the ('\n') is not required for join. When you call 'lines', it still has the '\n' character in the string, and when you join, it defaults to join without a delimiter, so it's putting them back together with the newline still there.

> File.write ARGV[1], (IO.readlines ARGV[0]).sort.join('\n')

not valid Ruby?

  File.open(ARGV[1], 'w') { |file| file.write(File.read(ARGV[0]).lines.sort.join) }

http://stackoverflow.com/questions/8278680/ruby-undefined-me...

Or:

   ruby -e 'puts ARGF.sort' file.txt

Also, this reads stdin if the argument is omitted.

> Would be great if somebody could post Python/Ruby/.. code here that achieves the same, just to compare.

It's actually really hard to completely rewrite that program in most scripting languages: they tend not to have the same concept of undefined behaviour.

'sys.argv[2]' in python (with sys.argv = ['thing.py']) is fully defined (raises an IndexError). 'argv[2]' in C++ (with argv = (char*[2]) { "./thing", 0 }) is undefined.

If you scripting language has an FFI library (ctypes in python, say) then you could probably do something equivalent.

good point about the UB. Together with shin_lao's comment about the error/exception handling bascially sums up why the author's claim only makes some sense, but not a lot.

Defined behavior is a subset of undefined behavior. Specifically, "raises an IndexError" is a perfectly legal thing for a C++ implementation to do in this case. Approximately zero real world implementations do this, but nothing says they can't.

Python:

    import sys

    data = open(sys.argv[1]).readlines()
    data.sort()
    open(sys.argv[2],'w').write(''.join(data))

Not bad, however with the ''.join(data) you're effectively doubling the amount of memory needed for large files, because it will build the entire output string in memory (and you've already got the list in memory). Better to use writelines(). You can also use sorted() to iterate over and sort the input lines automatically:

   sorted_lines = sorted(open(sys.argv[1]))
   open(sys.argv[2]).writelines(sorted_lines)

PowerShell:

  function Sort-File ($Path, $Destination) {
    Get-Content $Path | Sort-Object | Out-File $Destination
  }

Granted, shells are probably at an advantage when it comes to file management. I'd think a bash variant would be quite short as well.

Aiming for succinctness by using aliases and using redirection instead of cmdlets for writing the file (shorter, but less flexible):

  function sf { gc $args[0] | sort > $args[1] }

That's not code anyone should write outside code golfing, though ;-)

Yeah, the bash variant is obscenely short as well. Short enough that it isn't worth having as a separate program.

    cat in.txt | sort > out.txt

No need for cat!

    sort in.txt > out.txt

Curious, in case I'm missing something. Is what you've written equivalent to the following, or is there a reason for the pipe?

sort in.txt > out.txt

From the other comment, it is completely equivalent, and I will admit that the pipe is due purely to my ignorance. I usually use `sort` for sorting the output of other commands, and so I forget that it can open a file on its own.

That usage of cat is "wrong" in general:

    cat file | command > out

is the same as

    command < file > out

Don't worry about the cat it does have a purpose: portability. cat in.txt | sort > out.txt works in powershell(cat is an alias for Get-Content and sort is an alias for sort-object)(without the sort it wouln't work).

For those that don't know: there is/was an ancient meme called the "Useless Use of Cat Award" about that subject. See for instance http://en.wikipedia.org/wiki/Cat_(Unix)#Useless_use_of_cat, http://partmaps.org/era/unix/award.html, or http://unix.stackexchange.com/questions/16279/should-i-care-...

We are talking about a script, so it should be:

cat $1 | sort > $2

PHP:

    <?
    $lines=file($argv[1]);
    sort($lines);
    file_put_contents($argv[2],implode("\n",$lines));

Unix shell:

    sort

Another way to do it in C++ is:

  #include <iterator>
  #include<vector>
  #include<algorithm>
  #include<fstream>

  using namespace std; // Bad but shorter!

  int main(int argc, char** argv) {
      ifstream infile(argv[1]);
      ofstream outfile(argv[2]);
      ifstream_iterator<string> start(infile), end;
      vector<string> lines(start, end);
      sort(lines.begin(), lines.end());
      copy(lines.begin(), lines.end(), ostream_iterator<char>(outfile, "\n"));
  }

>Small nitpcik to the author

ifile, line and data have to be defined before the loop. ofile could be defined closer to usage, but I think it's much clearer where it is, grouping in and out defs together making it obvious how argv is used for each.

In Rebol:

  Rebol []

  files: map-each n system/options/args [to-file n]
  write/lines files/2 sort read/lines files/1

Python (assuming that the last line ends with a newline):

  import sys
  open(sys.argv[2], "w").writelines(sorted(open(sys.argv[1])))

Reply on news.ycombinator.com