Question

I am unfortunately having to use python with version before 2.4 So I don’t have the sorted built-in function. I need to sort a list; I found it is not possibly to go with

for aName in mylist.sort(lambda x,y:cmp(x.getName,y.getName)):

because then I would get an error saying

TypeError: iteration over non-sequence

I want to do the following:

for aName in sorted(mylist,key=lambda x:x.getName):

Can anybody help me with this? Many thanks.

Answer 1

You said that you wanted to do for aName in sorted(mylist,key=lambda x:x.getName): but accepted an answer that did something else …

It’s quite possible to write code that will run on multiple versions of Python 2.x. Here’s how to retrofit sorted() functionality to Python 2.1 to 2.3. It uses the DSU (decorate-sort-undecorate) aka “Schwartzian Transform” method … see this section of the Sorting HOWTO but do read the whole HOWTO; it’s very informative.

try:
    sorted
    def mysorted(iterable, key, reverse=0):
        return sorted(iterable, key=key, reverse=reverse)
except NameError: # doesn't have "sorted"
    def mysorted(iterable, key, reverse=0):
        temp = [(key(x), x) for x in iterable]
        temp.sort()
        if reverse:
            return [temp[i][1] for i in xrange(len(temp) - 1, -1, -1)]
        return [t[1] for t in temp]

mylist = 'tom dick harriet alfred zechariah'.split()
mykey = lambda x: x[2:] # ignore 1st 2 characters
print mylist
print mysorted(mylist, mykey)
print mysorted(mylist, mykey, reverse=1)

Running the above script with Python 2.7.1 and 2.1.3 produces the same output:

['tom', 'dick', 'harriet', 'alfred', 'zechariah']
['zechariah', 'dick', 'alfred', 'tom', 'harriet']
['harriet', 'tom', 'alfred', 'dick', 'zechariah']