Collections In Python | Introduction To Python Collections

Python programming language has four collection data types- list, tuple, sets and dictionary. But python also comes with a built-in module known as collections which has specialized data structures which basically covers for the shortcomings of the four data types. In this blog, we will go through each of those specialized data structures in detail.

  What Are Collections In Python?

Collections in python are basically container data types, namely lists, sets, tuples, dictionary. They have different characteristics based on the declaration and the usage.

A list is declared in square brackets, it is mutable, stores duplicate values and elements can be accessed using indexes.

A tuple is ordered and immutable in nature, although duplicate entries can be there inside a tuple.

A set is unordered and declared in square brackets. It is not indexed and does not have duplicate entries as well.

A dictionary has key value pairs and is mutable in nature. We use square brackets to declare a dictionary.

These are the python’s general purpose built-in container data types. But as we all know, python always has a little something extra to offer. It comes with a python module named collections which has specialized data structures.

Specialized Collection Data Structures

Collections module in python implements specialized data structures which provide alternative to python’s built-in container data types. Following are the specialized data structures in collections module.

1. namedtuple( )
2. deque
3. Chainmap
4. Counter
5. OrderedDict
6. defaultdict
7. UserDict
8. UserList
9. UserString

namedtuple( )

It returns a tuple with a named entry, which means there will be a name assigned to each value in the tuple. It overcomes the problem of accessing the elements using the index values. With namedtuple( ) it becomes easier to access these values, since you do not have to remember the index values to get specific elements.

How It Works?

First of all, you must import collections module, it does not require installation.

from collections import namedtuple

Look at the following code to understand how you can use namedtuple.

a = namedtuple('courses' , 'name , tech')
s = a('data science' , 'python')
print(s)

#the output will be courses(name='python' , tech='python')

How To Create A namedtuple Using A List?

s._make(['data science' , 'python'])
#the output will be same as before.

deque

deque pronounced as ‘deck’ is an optimized list to perform insertion and deletion easily.

How it works?

#creating a deque
from collections import deque

a = ['d' , 'u' , 'r' , 'e' , 'k']
a1 = deque(a)
print(a1)
#the output will be deque(['d' , 'u' , 'r' , 'e' , 'k'])

Now lets take a look at how we will insert and remove items from deque.

a1.append('a')
print(a1)
# the output will be deque(['d' , 'u' , 'r' , 'e' , 'k' , 'a'])
a1.appendleft('e')
print(a1)
# the output will be deque(['e' , 'd' , 'u' , 'r' , 'e' , 'k' , 'a'])

As should be obvious, inserting a component is enhanced utilizing deque, also you can remove components as well.

a1.pop()
print(a1)
#the output will be deque(['e' , 'd' , 'u' , 'r' , 'e' , 'k'])
a1.popleft()
print(a1)
#the output will be deque(['d' , 'u' , 'r' , 'e' , 'k'])

Similar to the built-in data types, there are several other operations that we can perform on a deque. Like counting elements or clearing the deque etc.

ChainMap

It is a dictionary like class which is able to make a single view of multiple mappings. It basically returns a list of several other dictionaries. Suppose you have two dictionaries with several key value pairs, in this case ChainMap will make a single list with both the dictionaries in it.

How it works?

from collections import ChainMap
a = { 1: 'edureka' , 2: 'python'}
b = {3: 'data science' , 4: 'Machine learning'}
c = ChainMap(a,b)
print(c)
#the output will be ChainMap[{1: 'edureka' , 2: 'python'} , {3: 'data science' , 4: 'Machine learning'}]

To access or insert elements we use the keys as index. But to add a new dictionary in the ChainMap we use the following approach.

a1 = { 5: 'AI' , 6: 'neural networks'}
c1 = c.new_child(a1)
print(c1)
#the output will be ChainMap[{1: 'edureka' , 2: 'python'} , {3: 'data science' , 4: 'Machine learning'}, { 5: 'AI' , 6: 'neural networks'}]

Counter

It is a dictionary subclass which is used to count hashable objects.

How it works?

from collections import Counter
a = [1,1,1,1,2,3,3,4,3,3,4]
c = Counter(a)
print(c)
#the output will be Counter = ({1:4 , 2:1 , 3:4 , 4:2})

In addition to the operations you can perform on a dictionary Counter has 3 more operations that we can perform.

1. element function – It returns a list containing all the elements in the Counter.
2. Most_common( ) – It returns a sorted list with the count of each element in the Counter.
3. Subtract( ) – It takes an iterable object as an argument and deducts the count of the elements in the Counter.

OrderedDict

It is a dictionary subclass which remembers the order in which the entries were added. Basically, even if you change the value of the key, the position will not be changed because of the order in which it was inserted in the dictionary.

How it works?

from collections import OrderedDict
od = OrderedDict()
od[1] = 'e'
od[2] = 'd'
od[3] = 'u'
od[4] = 'r'
od[5] = 'e'
od[6] = 'k'
od[7] = 'a'
print(od)
#the output will be OrderedDict[(1 , 'e'), (2 , 'd') , (3 , 'u'), (4 , 'r'), (5 , 'e'), (6 , 'k'), (7 , 'a')]

It does not matter what value gets inserted in the dictionary, the OrderedDict remembers the order in which it was inserted and gets the output accordingly. Even if we change the value of the key. Lets say, if we change the key value 4 to 8, the order will not change in the output.

defaultdict

It is a dictionary subclass which calls a factory function to supply missing values. In general, it does not throw any errors when a missing key value is called in a dictionary.

How it works?

from collections import defaultdict
d = defaultdict(int)
#we have to specify a type as well.
d[1] = 'edureka'
d[2] = 'python'
print(d[3])
#it will give the output as 0 instead of keyerror.

UserDict

This class acts as a wrapper around dictionary objects. The need for this class came from the necessity to subclass directly from dict. It becomes easier to work with this class as the underlying dictionary becomes an attribute.

class collections.UserDict([initialdata])

This class simulates a dictionary. The content of the instance are kept in a regular dictionary which can be accessed with the ‘data’ attribute of the class UserDict. The reference of initial data is not kept, for it to be used for other purposes.

UserList

This class acts like a wrapper around the list objects. It is a useful base class for other list like classes which can inherit from them and override the existing methods or even add a fewer new ones as well.

The need for this class came from the necessity to subclass directly from list. It becomes easier to work with this class as the underlying list becomes an attribute.

class collections.UserList([list])

It is the class that simulates a list. The contents of the instance are kept in a customary list. The sub-classes of the list are relied upon to offer a constructor which can be called with either no or one contention.

In this blog, we have learnt about the specialized data structures that comes with collections module in python. Optimization leads to better performance and enhanced results.

Have any queries? mention them in comments, and we will get back to you as soon as possible.