Question

In Python 2.2 (don't ask), what's the neatest way to sort a list and remove duplicates?

I can obviously write a function that would sort() then iterate, but am wondering if there's an idiomatic one-liner.

edit: The list is short, so efficiency is not a concern. Also, the elements are immutable.

Was it helpful?

Solution

For old python versions, and since you're using strings, there's no one-liner I can think of, but a pattern would probably be this, using dictionaries:

def sorted_uniq(your_list):
    table = {}
    for s in your_list:
        table[s] = None
    k = table.keys()
    k.sort()
    return k

Adapted from an ancient ActiveState code snippet thread that Alex Martelli himself wrote several comments on: http://code.activestate.com/recipes/52560/

A shorter way with list comprehensions:

def sort_uniq(alist):
   d = {}
   mod_list = [d.setdefault(i,i) for i in alist if i not in d]
   mod_list.sort()
   return mod_list

Aside from Steven's neat (yet slightly unattractive) one liner, I think this heads toward the fewest lines and most idiomatic way of doing it with Python 2.2:

Thanks to Steven Rumbalski in the comments, the 2nd version can be condensed further with python's zip function:

def sort_uniq(alist):
   mod_list = dict(zip(alist,alist)).keys()
   mod_list.sort()
   return mod_list

If list.sort() didn't operate by side effect, we'd have a one liner. ;)

OTHER TIPS

Idiomatic and a one-liner? No.

Here's a non-idiomatic butt-ugly one-liner.

>>> x = [4, 3, 3, 2, 4, 1]
>>> [y for y in (locals().__setitem__('d',{}) or x.sort() or x) 
        if y not in d and (d.__setitem__(y, None) or True)]
[1, 2, 3, 4]

If a simple two-liner is acceptable:

x = [4, 3, 3, 2, 4, 1]
x = dict(map(None,x,[])).keys()
x.sort()

Or make two small helper functions (works for any sequence):

def unique(it):
    return dict(map(None,it,[])).keys()

def sorted(it):
    alist = [item for item in it]
    alist.sort()
    return alist

print sorted(unique([4, 3, 3, 2, 4, 1]))

gives

[1, 2, 3, 4]

And finally, a semipythonic one liner:

x = [4, 3, 3, 2, 4, 1]
x.sort() or [s for s, t in zip(x, x[1:] + [None]) if s != t]

For the record, Python 2.2 does have sets, but under the "sets" module, so this will get you a long way:

from sets import Set
myList = list(Set(myList))
# now we're duplicate-free, a standard sorting might be enough
myList.sort()

Probably the best answer is to use a binary tree:

# Make yield work in Python 2.2
from __future__ import generators

class TreeNode(object):
    def __init__(self, value):
        self.left = None
        self.right = None
        self.value = value

    def add(self, value):
        if value == self.value:
            return
        if value < self.value:
            if self.left is None:
                self.left = TreeNode(value)
            else:
                self.left.add(value)
        else:
            if self.right is None:
                self.right = TreeNode(value)
            else:
                self.right.add(value)

    def __iter__(self):
        if self.left is not None:
            for value in self.left:
                yield value
        yield self.value
        if self.right is not None:
            for value in self.right:
                yield value

class DedupeSorter(object):
    def __init__(self):
        self.root = None

    def add(self, value):
        if self.root is None:
            self.root = TreeNode(value)
        else:
            self.root.add(value)

    def __iter__(self):
        if self.root is None:
            return []
        else:
            return self.root.__iter__()

def dedupe_and_sort(l):
    sorter = DedupeSorter()
    for value in l:
        sorter.add(value)
    return list(sorter)

Definitely not idiomatic, but should be fast. It basically creates a tree-based set and iterates over it. I don't have Python 2.2 so hopefully it works :p

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top