What are Python Itertools Functions?


According to Python’s documentation, Itertools is a Python module that gives a set of quick and memory-efficient instruments for working with Python iterators. These instruments can be utilized by themselves or together, they usually make it potential to succinctly and effectively create and work with iterators in a quick and memory-efficient method.

The Itertools module comprises features that make it simpler to work with iterators, notably when dealing with giant units of information. Itertools features can work on present iterators to create much more complicated Python iterators.

Additionally, Itertools will help builders cut back errors when working with iterators and write cleaner, readable, and maintainable code.


Based on the performance that the iterators within the Itertools module present, they are often categorized into the next sorts:

#1. Infinite Iterators

These are iterators that will let you work with infinite sequences and run a loop infinitely if there’s no situation put in to interrupt out of the loop. Such iterators are helpful when simulating infinite loops or producing an unbounded sequence. Itertools has three infinite iterators, which embrace rely(), cycle(), and repeat().

#2. Combinatoric Iterators

Combinatoric Iterators comprise features that can be utilized to work on cartesian merchandise and carry out mixtures and permutations of components contained inside an iterable. These are the go-to features when looking for all potential methods to rearrange or mix components in an iterable. Itertools has 4 combinatoric iterators. These are product(), permutations(), mixtures() and combinations_with_replacement().

#3. Iterators Terminating On The Shortest Input Sequence

These are terminating iterators that are used on finite sequences and generate an output based mostly on the kind of operate used. Examples of those terminating iterators embrace: accumulate(), chain(), chain.from_iterable(), compress(), dropwhile(), filterfalse(), groupby(), islice(), pairwise(), starmap(), takewhile(), tee(), and zip_longest().


Let us take a look at how completely different Itertools features work in keeping with their sort:

Infinite Iterators

The three infinite iterators embrace:

#1. rely()

The rely(begin, step) operate generates an infinite sequence of numbers ranging from the beginning worth. The operate takes two non-compulsory arguments: begin and step. The argument begin units the place the sequence of numbers ought to begin. By default, it begins at 0 if a begin worth just isn’t offered. step units the distinction between every consecutive quantity. The default step worth is 1.

import itertools
# rely beginning at 4, making steps of two  
for i in itertools.rely(4, 2):
    # situation to finish the loop avoiding infinite looping
    if i == 14:
        print(i) # output - 4, 6, 8, 10, 12



#2. cycle()

cycle(iterable) operate takes an iterable as an argument after which cycles by way of the iterable permitting entry to gadgets within the iterable within the order they seem.

For occasion, if we cross in [“red”, “green”, “yellow”] into cycle(), within the first cycle, we’ll have entry to “red”; within the second cycle we’ll have entry to “green”, then “yellow”. In the fourth cycle, since all components have been exhausted within the iterable, we’ll begin over at “red” after which go on infinitely.

When calling cycle() you retailer its end in a variable to create an iterator that maintains its state. This ensures the cycle doesn’t begin throughout each time, supplying you with entry to solely the primary aspect.

import itertools

colours = ["red", "green", "yellow"]
# cross in colours into cycle()
color_cycle = itertools.cycle(colours)

# vary used to cease the infinite loop as soon as we have printed 7 occasions
# subsequent() used to return the following merchandise from the iterator
for i in vary(7):



#3. repeat()

repeat(elem,n) takes two arguments, a component to repeat (elem), and the variety of occasions you wish to repeat the aspect(n). The aspect you wish to repeat is usually a single worth or an iterable. If you don’t cross in, n, the aspect shall be repeated infinitely.

import itertools
for i in itertools.repeat(10, 3):



Combinatoric Iterators

The combinatoric iterators embrace:

#1. product()

product() is a operate used to compute the cartesian product of the iterable handed to it. If we have now two iterables or units, for instance, x = {7,8} and y = {1,2,3}, the cartesian product of x and y will include all potential mixtures of components from x and y, the place the primary aspect is from x and the second from y. The cartesian product of x and y on this case is [(7, 1), (7, 2), (7, 3), (8, 1), (8, 2), (8, 3)].

product() takes an non-compulsory parameter known as repeat which is used to compute the cartesian product of an iterable with itself. repeat specifies the variety of repetitions for every aspect from the enter iterables when computing the Cartesian product.

For occasion, calling product(‘ABCD’, repeat=2) yields mixtures corresponding to (‘A’, ‘A’), (‘A’, ‘B’), (‘A’, ‘C’), and so forth. If repeat was set to three, the operate would yield mixtures corresponding to (‘A’, ‘A’, ‘A’), (‘A’, ‘A’, ‘B’), (‘A’, ‘A’, ‘C’), (‘A’, ‘A’, ‘D’) and so forth.

from itertools import product
# product() with the non-compulsory repeat argument
print("product() with the optional repeat argument ")
print(checklist(product('ABC', repeat = 2)))

# product with no repeat
print("product() WITHOUT an optional repeat argument")
print(checklist(product([7,8], [1,2,3])))


product() with the non-compulsory repeat argument 
[('A', 'A'), ('A', 'B'), ('A', 'C'), ('B', 'A'), ('B', 'B'), ('B', 'C'), ('C', 'A'), ('C', 'B'), ('C', 'C')]
product() WITHOUT an non-compulsory repeat argument
[(7, 1), (7, 2), (7, 3), (8, 1), (8, 2), (8, 3)]

#2. permutations()

permutations(iterable, group_size) returns all potential permutations of the iterable handed into it. A permutation represents the variety of methods components in a set will be ordered. permutations() takes an non-compulsory argument group_size. If group_size just isn’t specified, the permutations generated would be the identical dimension because the size of the iterable handed into the operate

import itertools
numbers = [1, 2, 3]
sized_permutations = checklist(itertools.permutations(numbers,2))
unsized_permuatations = checklist(itertools.permutations(numbers))

print("Permutations with a size of 2")
print("Permutations with NO size argument")


Permutations with a bunch dimension of two
[(1, 2), (1, 3), (2, 1), (2, 3), (3, 1), (3, 2)]
Permutations with NO dimension argument
[(1, 2, 3), (1, 3, 2), (2, 1, 3), (2, 3, 1), (3, 1, 2), (3, 2, 1)]

#3. mixtures()

mixtures(iterable, dimension) returns all potential mixtures of an iterable of a given size from the weather within the iterable handed into the operate. The dimension argument specifies the dimensions of every mixture.

The outcomes are ordered. Combination differs barely from permutations. With permutation, the order issues, however with mixture, the order doesn’t matter. For occasion, in [A, B, C] there are 6 permutations: AB, AC, BA, BC, CA, CB however solely 3 mixtures AB, AC, BC.

import itertools
numbers = [1, 2, 3,4]
size2_combination = checklist(itertools.mixtures(numbers,2))
size3_combination = checklist(itertools.mixtures(numbers, 3))

print("Combinations with a size of 2")
print("Combinations with a size of 3")


Combinations with a dimension of two
[(1, 2), (1, 3), (1, 4), (2, 3), (2, 4), (3, 4)]
Combinations with a dimension of three
[(1, 2, 3), (1, 2, 4), (1, 3, 4), (2, 3, 4)]

#4. combinations_with_replacement()

combinations_with_replacement(iterable, dimension) generates all potential mixtures of an iterable of a given size from the iterable handed into the operate and permits for repeated components within the output mixtures. The dimension determines the dimensions of the mixtures generated.

This operate differs from mixtures() in that it offers mixtures the place a component will be repeated greater than as soon as. For occasion, you will get a mixture corresponding to (1,1) which you’ll be able to’t with mixture().

import itertools
numbers = [1, 2, 3,4]

size2_combination = checklist(itertools.combinations_with_replacement(numbers,2))
print("Combinations_with_replacement => size 2")


Combinations_with_replacement => dimension 2
[(1, 1), (1, 2), (1, 3), (1, 4), (2, 2), (2, 3), (2, 4), (3, 3), (3, 4), (4, 4)]

Terminating iterators

This contains iterators corresponding to:

#1. accumulate()

accumulate(iterable, operate) takes an iterable and a second non-compulsory argument which is a operate. It then returns the amassed results of making use of the operate in every iteration on components on the iterable. If no operate is handed, addition is completed and the amassed outcomes are returned.

import itertools
import operator
numbers = [1, 2, 3, 4, 5]

# Accumulate the sum of numbers
accumulated_val = itertools.accumulate(numbers)
accumulated_mul = itertools.accumulate(numbers, operator.mul)
print("Accumulate with no function")
print("Accumulate with multiplication")


Accumulate with no operate
[1, 3, 6, 10, 15]
Accumulate with multiplication
[1, 2, 6, 24, 120]

#2. chain()

chain(iterable_1, iterable_2, …) takes a number of iterables and chains them collectively producing a single iterable containing values from the iterables handed to the chain() operate

import itertools

letters = ['A', 'B', 'C', 'D']
numbers = [1, 2, 3]
colours = ['red', 'green', 'yellow']

# Chain letters and numbers collectively
chained_iterable = checklist(itertools.chain(letters, numbers, colours))


['A', 'B', 'C', 'D', 1, 2, 3, 'red', 'green', 'yellow']

#3. chain.from_iterable()

chain.from_iterable(iterable)  this operate is much like chain(). However, it differs from the chain in that it solely takes a single iterable containing sub-iterables and chains them collectively.

import itertools

letters = ['A', 'B', 'C', 'D']
numbers = [1, 2, 3]
colours = ['red', 'green', 'yellow']

iterable = ['hello',colors, letters, numbers]
chain = checklist(itertools.chain.from_iterable(iterable))


['h', 'e', 'l', 'l', 'o', 'red', 'green', 'yellow', 'A', 'B', 'C', 'D', 1, 2, 3]

#4. compress()

compress(information, selectors) takes in two arguments, information which is an iterable, and selectors which is an iterable containing booleans values true, and false. 1, 0 can be used as alternate options to the boolean values true and false. compress() then filters the handed information utilizing the corresponding components handed within the selector.

Values in information that correspond to the worth true or 1 within the selector are chosen, whereas the remainder which correspond to false or 0 are ignored. If you cross fewer booleans in selectors than the variety of gadgets in information all the weather past the handed booleans in selectors are ignored

import itertools

# information has 10 gadgets
information = ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J']
# passing in 9 selector gadgets
selectors = [True, False, 1, False, 0, 1, True, False, 1]

# Select components from information based mostly on selectors
filtered_data = checklist(itertools.compress(information, selectors))


['A', 'C', 'F', 'G', 'I']

#5. dropwhile()

dropwhile(operate, sequence) takes in a operate with the situation that returns true or false and a sequence of values. It then drops all values till the situation handed returns False. Once the situation returns false, the remainder of the weather are included in its outcomes no matter whether or not they’d return True or False.

import itertools

numbers = [1, 2, 3, 4, 5, 1, 6, 7, 2, 1, 8, 9, 0, 7]

# Drop components till the handed situation is False
filtered_numbers = checklist(itertools.dropwhile(lambda x: x < 5, numbers))


[5, 1, 6, 7, 2, 1, 8, 9, 0, 7]

#6. filterfalse()

filterfalse(operate, sequence) takes in a operate, with a situation that evaluates to true or false and a sequence. It then returns values from the sequence which don’t fulfill the situation within the operate. 

import itertools

numbers = [1, 2, 3, 4, 2, 3 5, 6, 5, 8, 1, 2, 3, 6, 2, 7, 4, 3]

# Filter components for which situation is False
filtered_numbers = checklist(itertools.filterfalse(lambda x: x < 4, numbers))


[4, 5, 6, 5, 8, 6, 7, 4]

#7. groupby()

groupby(iterable, key) takes in an iterable and a key, then makes an iterator that returns consecutive keys and teams. For it to work, the iterable handed to it must be sorted on the identical key operate. The key operate pc a key worth for every aspect within the iterable.

import itertools

input_list = [("Domestic", "Cow"), ("Domestic", "Dog"), ("Domestic", "Cat"),("Wild", "Lion"), ("Wild", "Zebra"), ("Wild", "Elephant")]
classification = itertools.groupby(input_list,lambda x: x[0])
for key,worth in classification:


Domestic : [('Domestic', 'Cow'), ('Domestic', 'Dog'), ('Domestic', 'Cat')]
Wild : [('Wild', 'Lion'), ('Wild', 'Zebra'), ('Wild', 'Elephant')]

#8. islice()

islice(iterable, begin, cease, step) means that you can slice an iterable utilizing the begin, cease, and step values handed. The step argument is non-compulsory. Counting begins from 0 and the merchandise on the cease quantity just isn’t included.

import itertools

numbers = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18]

# Select components inside a variety
selected_numbers = checklist(itertools.islice(numbers, 2, 10))
selected_numbers_step= checklist(itertools.islice(numbers, 2, 10,2))
print("islice without setting a step value")
print("islice with a step value of 2")


islice with out setting a step worth
[3, 4, 5, 6, 7, 8, 9, 10]
islice with a step worth of two
[3, 5, 7, 9]

#9. pairwise()

pairwise(iterable) returns successive overlapping pairs taken from the iterable handed to it within the order they seem within the iterable. If the iterable handed to it has lower than two values, the outcome from pairwise() shall be empty.

from itertools import pairwise

numbers = [1, 2, 3, 4, 5, 6, 7, 8]
phrase = 'WORLD'
single = ['A']



[(1, 2), (2, 3), (3, 4), (4, 5), (5, 6), (6, 7), (7, 8)]
[('W', 'O'), ('O', 'R'), ('R', 'L'), ('L', 'D')]

#10. starmap()

starmap(operate, iterable) is a operate used as an alternative of map() when argument parameters are already grouped in tuples. startmap() applies a operate to the weather of the iterable handed to it. The iterable ought to have components grouped in tuples.

import itertools

iter_starmap = [(123, 63, 13), (5, 6, 52), (824, 51, 9), (26, 24, 16), (14, 15, 11)]
print (checklist(itertools.starmap(min, iter_starmap)))


[13, 5, 9, 16, 11]

#11. takewhile()

takewhile(operate, iterable) works within the reverse approach to dropwhile(). takewhile() takes in a operate with a situation to be evaluated and an iterable. It then contains all components within the iterable which fulfill the situation within the operate till False is returned. Once False is returned, all the next components within the iterable are ignored.

import itertools

numbers = [1, 2, 3, 4, 5, 1, 6, 7, 2, 1, 8, 9, 0, 7]

# Drop components till the handed situation is False
filtered_numbers = checklist(itertools.takewhile(lambda x: x < 5, numbers))


[1, 2, 3, 4]

#12. tee()

tee(iterable, n) takes in an iterable and returns a number of unbiased iterators. The variety of iterators to return is about by n, which by default is 2.

import itertools

numbers = [1, 2, 3, 4, 5]

# Create two unbiased iterators from numbers
iter1, iter2 = itertools.tee(numbers, 2)


[1, 2, 3, 4, 5]
[1, 2, 3, 4, 5]

#13. zip_longest()

zip_longest(iterables, fillvalue) takes in a number of iterators and a fillvalue. It then returns an iterator that aggregates components from every of the iterators handed to it. If the iterators are not of the identical size, the lacking values are changed by the fillvalue handed to the operate till the longest iterable has been exhausted.

import itertools

names = ['John', 'mathew', 'mary', 'Alice', 'Bob', 'Charlie', 'Fury']
ages = [25, 30, 12, 13, 42]

# Combine identify and ages, filling in lacking ages with a touch
mixed = itertools.zip_longest(names, ages, fillvalue="-")

for identify, age in mixed:
    print(identify, age)


John 25
mathew 30
mary 12
Alice 13
Bob 42
Charlie -
Fury -


Python itertools are an essential toolset for a Python developer. Python itertools are used extensively in purposeful programming, information processing, and transformation, information filtering and choice, grouping and aggregation, combining iterables, combinatorics, and when working with infinite sequences.

As a Python developer, you’ll profit significantly by studying about itertools so make certain to make use of this text to familiarize your self with Python Itertools.



Leave a Reply

Your email address will not be published. Required fields are marked *