Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Quicksort in 3 lines of shell (github.com/wyc)
31 points by wyc on Nov 11, 2014 | hide | past | favorite | 17 comments


Let me fix that for you.

    qsort(){
        local L="";
        local G="";
        [ $# -eq 1 ] && echo $1 && return;
        P=$1; 
        shift; 
        for i in $@; do 
            [ $i -lt $P ] && L="$L $i" || G="$G $i"; 
        done
        [ -z "$L" ] || L=`qsort $L`; 
        [ -z "$G" ] || G=`qsort $G`; 
        echo "$L $P $G"
    }
    qsort $@
Looks more like 14 lines to me.


Also: what? No median-of-three pivot selection?

  # these lines replace "P=$1" line:

  local first=$1
  eval local last=\${$#}
  eval local mid=\${$((($#+1)/2))}

  if [ $first -le $mid ] ; then
    if [ $mid -le $last ] ; then
      P=$mid
    elif [ $first -le $last ] ; then
      P=$last
    else
      P=$first
    fi
  else
    if [ $first -le $last ] ; then
      P=$first
    elif [ $mid -le $last ] ; then
      P=$last
    else
      P=$mid
    fi
  fi
That little enhancement blows up the LOC quite a bit. :)


Also, suppose we abstract the operators a little bit:

  if [ $x $LT $y ] ; ...
If we set LT to "-lt" we get numeric comparison.

Change it to "-ot" and we can order paths by modification timestamp.

Set LT to "<" and obtain lexicographic sorting on strings.



Well do be fair none of the commands in this quicksort require user input at least :)


> L=`qsort $L`

POSIX has standardized $(qsort $L) syntax since 1990. It's okay to use it now.


The time complexity looks very bad (assuming I can read shell, which I am not sure about).

First it appears it has the classic bug of passing items equal to the partition to sub-sorts (G in this case). This guarantees N recurses (and at least c N^2 operations) when trying to sort an array that is all duplicates of a single value. Also the list concatenations ( L="$L $i" and G="$G $i") which often make mere list assembly take c N^2 time (pretty much forced by the string representations). So this could be a c N^3 worst-case sorting algorithm (possibly worse than bubble sort).

One of my articles on the pain of quicksort: http://www.win-vector.com/blog/2008/04/sorting-in-anger/


Haha, there are few places where this submission would result in such a decent analysis. Thanks.


Maybe we should all learn to always prepend "Pseudo" to these "quicksorts"?


Is it not really a quicksort?


I meant that the canonical quicksort algorithm which Hoare came up uses in-place partitioning. Probably you can still call "quicksort" to these nice snippets (be it in Prolog, Erlang or Haskell), but mentioning the fact about the deviation from the canonical implementation could be useful for readers.


It's a bad implementation of it, see jmount's comment.


I guess he meant pseudo 3 lines.


A couple of characters less in C ( I didn't count whitespace in both ):

void qs( int* a , int z )

{

if(z > 1){ ; int p = a[z/2], s = 0, e = z-1, t; while(s <= e) {

for(;a[s] < p; s++); for(;a[e] > p; e--); if(s <= e) { t = a[s];

a[s++] = a[e]; a[e--] = t; } } qs( a , e+1 ); qs( a+s , z-s ); }

}


https://gist.github.com/pdkl95/e97749edc3ba58825a16#file-bas...

I what must be a lapse of sanity, I wrote mergesort. It is quite a bit more than 3 - err, 14 - lines, but it does include an option to see view the array nesting.

It uses slightly better array management... and then throws out any gains by rebuilding a new array every time an element is shift off the front.

// for when you really don't want to use sort(1)


It's not 3 commands so I'm sure you can get it down to 1 line.


Hey, check out this 1-liner HTTP server!

:%s/\n/ /g




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: