DEV Community

Micheal Ojemoron
Micheal Ojemoron

Posted on

Pythonic way to aggregate or group elements in a list using dict.get and dict.setdefault

Using dict.get

How often have you aggregated an item by its group like this:

ar= [1,1,3,1,2,1,3,3,3,3]
pairs = {}
for i in ar:
    if i not in pairs:
         pairs[i] = 0
    pairs[i] = pairs[i] + 1


print(pairs)
{1: 4, 3: 5, 2: 1}
Enter fullscreen mode Exit fullscreen mode

The above code shows how to get the frequency of elements in a list.
While this code is good, the time complexity of in(membership operator) is O(n) worse case.
We can make this code better and more pythonic by using python's dict.get(k,d) method.

Using the same list, we have:

pairs = {}
for i in ar:
    pairs[i] = pairs.get(i, 0) + 1
print(pairs)
{1: 4, 3: 5, 2: 1}
Enter fullscreen mode Exit fullscreen mode

the trick here is that dict.get(k,d) returns None by default or the specified default value if a key doesn't exist.

Using dict.setdefault

Similarly when grouping things together we usually do this:

items_by_type = {}
for item in items:
    if item.type not in items_by_type:
        items_by_type[item.type] = list()
    items_by_type[item.type].append(item)
Enter fullscreen mode Exit fullscreen mode

Let's make it better and more pythonic:

items_by_type = {}
for item in items:
    items_by_type.setdefault(item.type, list()).append(item)
Enter fullscreen mode Exit fullscreen mode

The trick here is that setdefault(k,d) only sets the item if it does not
exist. If it does exist then it simply returns the item.

Final Thoughts

You can use the Counter class from the collection module to get the frequency of elements in a list.
Just do this:

from collections import Counter
#using our above example:
print(Counter(ar))
Enter fullscreen mode Exit fullscreen mode

Happy coding! โœŒ

Please drop your comments :)

Top comments (10)

Collapse
 
patarapolw profile image
Pacharapol Withayasakpunt

This is just too magical, and hard to understand.

items_by_type = {}
for item in items:
    items_by_type.setdefault(item.type, list()).append(item)

A less magical version is defaultdict (from collections)

items_by_type = defaultdict(list)
for item in items:
    items_by_type[item.type].append(item)
Collapse
 
waylonwalker profile image
Waylon Walker • Edited

Diddo the swap from

for i in range(len(items)):
   ...

to

for item in items:
   ...
Collapse
 
mojemoron profile image
Micheal Ojemoron

you are absolutely right, I have updated the code. Thanks

Collapse
 
mojemoron profile image
Micheal Ojemoron • Edited

You are right๐Ÿ˜, fortunately there are several ways to do things in Python. I am more concerned about using dictionary methods. Thank you

Collapse
 
waylonwalker profile image
Waylon Walker

Good walk of the different dictionary methods.

Collapse
 
mojemoron profile image
Micheal Ojemoron

Thanks

Collapse
 
waylonwalker profile image
Waylon Walker

I love the Counter object. I've had a rest API getting bogged down with pandas trying to count something and it took a few seconds for each response. Counter was just a few ms.

Collapse
 
mojemoron profile image
Micheal Ojemoron

Great!

Collapse
 
bjoxiah profile image
Blessed J. Eikhiena

Great!

Collapse
 
mojemoron profile image
Micheal Ojemoron

Thanks