Loop problem while iterating through a list and removing recurring elements [duplicate]

This question already has answers here: Modifying list while iterating [duplicate] (7 answers) 开发者_运维百科 Closed 8 years ago.

I want to iterate through a list, and remove the items that count more than once, so they don't get printed repeatedly by the for loop.

However, some items appearing only one time in the list seem to get affected too by this, and I can't figure out why.

Any input would be greatly appreciated.

Example Output:

listy = [2,2,1,3,4,2,1,2,3,4,5]
for i in listy:
  if listy.count(i)>1:
    print i, listy.count(i)
    while i in listy: listy.remove(i)
  else:
    print i, listy.count(i)

Outputs:

 2 4
 3 2
 1 2

thus ignoring completely 4 and 5.

---

You should not modify a list while iterating over it. This one should work:

listy = [2,2,1,3,4,2,1,2,3,4,5]
found = set()
for i in listy:
    if not i in found:
        print i, listy.count(i)
        found.add(i)

The result is:

2 4
1 2
3 2
4 2
5 1

---

The reason for your problems is that you modify the list while you are iterating over it.

If you don't care about the order in which items appear in the output and don't care about the count, you can simply use use a set:

>>> listy = [2,2,1,3,4,2,1,2,3,4,5]
>>> print set(listy)
set([1, 2, 3, 4, 5])

If you do care about the count, use the Counter class from the collections module in the Standard Library:

>>> import collections
>>> collections.Counter(listy)
Counter({2: 4, 1: 2, 3: 2, 4: 2, 5: 1})
>>> c = collections.Counter(listy)
>>> for item in c.iteritems():
...     print "%i has a count of %i" % item
... 
1 has a count of 2
2 has a count of 4
3 has a count of 2
4 has a count of 2
5 has a count of 1

If you do care about both the order and the count, you have to build a second list:

>>> checked = []
>>> counts = []
>>> for item in listy: 
>>>     if item not in checked: 
>>>         checked.append(item) 
>>>         counts.append(listy.count(item))
>>> print zip(checked, counts)
... [(2, 4), (1, 2), (3, 2), (4, 2), (5, 1)]

This is the least efficient solution, of course.

If you don't want to keep the counts for later, you don't need the counts list:

listy = [2,2,1,3,4,2,1,2,3,4,5]
checked = set()
for item in listy: 
    # "continue early" looks better when there is lots of code for
    # handling the other case
    if item in checked:     
        continue

    checked.add(item) 
    print item, listy.count(item)

---

Don't modify a list while iterating over it, it will mess you up every time:

listy = [2,2,1,3,4,2,1,2,3,4,5]
#        *     *     * Get hit
for i in listy:
    print i
    if listy.count(i) > 1:
        print i, listy.count(i), 'item and occurences'
        while i in listy: listy.remove(i)
    else:
        print i, listy.count(i)

1. First, you remove four 2s. Two are right at the beginning, so that puts you at the first 1.
2. Then you advance one when you get the next i from listy, putting you at the first 3.
3. Then you remove two 3s. The first is right there, so that puts you at the first 4.
4. Then you advance one again. The 2 is gone already, so this puts you at the second 1.
5. You then delete both 1s; this moves you forward two spaces. The 2 and 3 are gone, so this puts you at the 5.
6. You advance one, this moves you off the end of the list so the loop is over.

If what you want is to print each item only once, you can use the simple set method, or you could use the itertools unique_everseen recipe:

def unique_everseen(iterable, key=None):
    "List unique elements, preserving order. Remember all elements ever seen."
    # unique_everseen('AAAABBBCCDAABBB') --> A B C D
    # unique_everseen('ABBCcAD', str.lower) --> A B C D
    seen = set()
    seen_add = seen.add
    if key is None:
        for element in ifilterfalse(seen.__contains__, iterable):
            seen_add(element)
            yield element
    else:
        for element in iterable:
            k = key(element)
            if k not in seen:
                seen_add(k)
                yield element

Which extends the basic set version to allow you to specify a special way to compare items.

If you want to know which items are only in the list once:

listy2 = filter(lambda i: listy.count(i) == 1, listy)

listy2 now has all the single occurrences.

If you don't like the lambda, just do:

def getsingles(listy):
    def singles(i):
        return listy.count(i) == 1
    return singles

then:

listy2 = filter(getsingles(listy), listy)

This makes a special function that will tell you which items are in listy only once.

---

The reason of the behavior you get is here, in the note:

http://docs.python.org/reference/compound_stmts.html#index-811

Update 1

agf's solution isn't a good one for performance reason: the list is filtered according to the count of each element. The counting is done for each element, that is to say the counting process that consists to run through the entire list to count, is done as many times as there are elements in list: it's overconsuming time, imagine if your list is 1000 length

A better solution I think is to use an instance of Counter:

import random
from collections import Counter

li = [ random.randint(0,20) for i in xrange(30)]

c = Counter(li)

print c
print type(c)

res = [ k for k in c if c[k]==1]
print res

result

Counter({8: 5, 0: 3, 4: 3, 9: 3, 2: 2, 5: 2, 11: 2, 3: 1, 6: 1, 10: 1, 12: 1, 15: 1, 16: 1, 17: 1, 18: 1, 19: 1, 20: 1})
<class 'collections.Counter'>
[3, 6, 10, 12, 15, 16, 17, 18, 19, 20]

Another solution would be to add the read elements in a set in order that the program avoids to make a count for an already seen element.

Update 2

errrr.... my solution is stupid, you don't want to select the element appearing only one time in the list....

Then the following code is the right one , I think:

import random
from collections import Counter

listy = [ random.randint(0,20) for i in xrange(30)]
print 'listy==',listy
print

c = Counter(listy)
print c
print type(c)
print

slimmed_listy = []
for el in listy:
    if el in c:
        slimmed_listy.append(el)
        print 'element',el,'  count ==',c[el]
        del c[el] 
print

print 'slimmed_listy==',slimmed_listy

result

listy== [13, 10, 1, 1, 13, 11, 18, 15, 3, 15, 12, 11, 15, 18, 11, 10, 14, 10, 20, 3, 18, 9, 11, 2, 19, 15, 5, 14, 1, 1]

Counter({1: 4, 11: 4, 15: 4, 10: 3, 18: 3, 3: 2, 13: 2, 14: 2, 2: 1, 5: 1, 9: 1, 12: 1, 19: 1, 20: 1})
<class 'collections.Counter'>

element 13   count == 2
element 10   count == 3
element 1   count == 4
element 11   count == 4
element 18   count == 3
element 15   count == 4
element 3   count == 2
element 12   count == 1
element 14   count == 2
element 20   count == 1
element 9   count == 1
element 2   count == 1
element 19   count == 1
element 5   count == 1

slimmed_listy== [13, 10, 1, 11, 18, 15, 3, 12, 14, 20, 9, 2, 19, 5]

In case you wouldn't want the result in the order of listy, the code would be even simpler

Update 3

If you want only to print, then I propose:

import random
from collections import Counter

listy = [ random.randint(0,20) for i in xrange(30)]
print 'listy==',listy
print


def gener(li):
    c = Counter(li)
    for el in li:
        if el in c:
            yield el,c[el]
            del c[el] 


print '\n'.join('element %4s   count %4s' % x for x in gener(listy))

result

listy== [16, 2, 4, 9, 15, 19, 1, 1, 3, 5, 12, 15, 12, 3, 17, 13, 8, 11, 4, 6, 15, 1, 0, 1, 3, 3, 6, 5, 0, 8]

element   16   count    1
element    2   count    1
element    4   count    2
element    9   count    1
element   15   count    3
element   19   count    1
element    1   count    4
element    3   count    4
element    5   count    2
element   12   count    2
element   17   count    1
element   13   count    1
element    8   count    2
element   11   count    1
element    6   count    2
element    0   count    2

---

Modifying a list while you iterate over it is a bad idea in every language I have encountered. My suggestion: don't do that. Here are some better ideas.

Use a set to find single occurrences

source = [2,2,1,3,4,2,1,2,3,4,5]
for s in set(source):
    print s

And you get this:

>>> source = [2,2,1,3,4,2,1,2,3,4,5]
>>> for s in set(source):
...     print s
... 
1
2
3
4
5

If you want the counts, use defaultdict

from collections import defaultdict
d = defaultdict(int)
source = [2,2,1,3,4,2,1,2,3,4,5]
for s in source:
    d[s] += 1

for k, v in d.iteritems():
    print k, v

You'll get this:

>>> for k, v in d.iteritems():
...     print k, v
... 
1 2
2 4
3 2
4 2
5 1

If you want your results sorted, use sort and operator

import operator
for k, v in sorted(d.iteritems(), key=operator.itemgetter(1)):
    print k, v

You'll get this:

>>> import operator
>>> for k, v in sorted(d.iteritems(), key=operator.itemgetter(1)):
...     print k, v
... 
5 1
1 2
3 2
4 2
2 4

---

I am not sure if it is a good idea to iterate the list and remove elements at the same time. If you really just want to output all items and their number of occurrences, I would do it like this:

listy = [2,2,1,3,4,2,1,2,3,4,5]
listx = []
listc = []
for i in listy:
    if not i in listx:
        listx += [i]
        listc += [listy.count(i)]
for x, c in zip(listx, listc):
    print x, c

---

Like agf said, modifying a list while you iterate it will cause problems. You could solve your code by using while and pop:

single_occurrences = []
while listy:
    i = listy.pop(0)
    count = listy.count(i)+1
    if count > 1:
        print i, count
        while i in listy: listy.remove(i)
    else:
        print i, count
    single_occurrences.append(i)

Output:

2 4
1 2
3 2
4 2
5 1

---

One way to do that would be to create a result list and test whether the tested value is in it :

res=[]
listy = [2,2,1,3,4,2,1,2,3,4,5]

for i in listy:
    if listy.count(i)>1 and i not in res:
        res.append(i)

for i in res:
    print i, listy.count(i)

Result :

2 4
1 2
3 2
4 2