Python Sets

Overview

Python sets are un-ordered. Care has to be taken to make sure you don’t assume sets to be ordered like lists are. In this sense they behave similar to dictionary which is also unordered.

Creating A Set

You create a new set in Python with the call set(), passing in a list of values you want in your set:

1

my_set = set([1, 2, 3, 4])

You can also create a set using set literal syntax:

1

my_set = { 1, 2, 3, 4 }

Set literal syntax looks very similar to dictionary literal syntax (which also uses {/}), except rather than key/value pairs there is just a single value. Set literal syntax is faster than set() call above as Python doesn’t have to create a list first.

Set Operations

Add

You can use the .add() function to add a single element to a set. If it already exists in the set, this function does nothing.

1
2
3
4

A = { 1, 2, 3 }
A.add(4)
print(A)
# stdout: {1, 2, 3, 4}

If you want to add multiple elements at the same time, see the .union() function below.

Union

A set union can be done either with .union() or |. It returns a set of all elements that are in either A or B:

1
2
3
4
5

A = { 1, 2, 3, 4 }
B = { 3, 4, 5, 6 }
C = A | B
print(C)
# stdout: {1, 2, 3, 4, 5, 6}

The union returns a new set and does not modify the input sets. To perform a union but to modify inplace (which can potentially save memory and increase performance), use update():

1
2
3
4

A = { 1, 2, 3, 4 }
B = { 3, 4, 5, 6 }
A.update(B)
print(A)

Intersection

A set intersection can be done with either .intersection() or &. It returns a set of only the elements which are in both A and B:

1
2
3
4
5

A = { 1, 2, 3, 4 }
B = { 3, 4, 5, 6 }
C = A & B
print(C)
# stdout: {3, 4}

To perform an inplace intersection:

1
2
3
4
5

A = { 1, 2, 3, 4 }
B = { 3, 4, 5, 6 }
A.intersection_update(B)
print(A)
# stdout: {3, 4}

Difference

The difference() function or - operator returns the difference between two sets (every item in A that is not also in B).

1
2
3
4

A = { 1, 2, 3, 4 }
B = { 3, 4, 5, 6 }
print(A - B)
# stdout: {1, 2}

To perform an inplace difference:

1
2
3
4
5

A = { 1, 2, 3, 4 }
B = { 3, 4, 5, 6 }
A.difference_update(B)
print(A)
# stdout: {1, 2}

Symmetric Difference

The .symmetric_difference() function or the ^ operator returns the elements belonging to either set A or set B, but not those that belong to both:

1
2
3
4
5

A = { 1, 2, 3, 4 }
B = { 3, 4, 5, 6 }
C = A ^ B
print(C)
# stdout: {1, 2, 5, 6}

To perform an inplace symmetric difference operation use symmetric_difference_update():

1
2
3
4
5

A = { 1, 2, 3, 4 }
B = { 3, 4, 5, 6 }
A.symmetric_difference_update(B)
print(A)
# stdout: {1, 2, 5, 6}

Subset

The .issubset() or <= operator returns true if A is a subset of B. A is a subset of B when B contains at least every element that A contains.

1
2
3
4
5
6
7
8
9

A = { 1, 2 }
B = { 1, 2, 3, 4}
print(A <= B)
# stdout: True

A = { 1, 2 }
B = { 2, 3, 4}
print(A <= B)
# stdout: False

Superset

The .issuperset() or >= operator returns True if A is a superset of B. A is a superset of B when A contains at least every element that B contains.

1
2
3
4
5
6
7
8
9

A = { 1, 2, 3 }
B = { 1, 2 }
print(A >= B)
# stdout: True

A = { 1, 2 }
B = { 2, 3, 4}
print(A >= B)
# stdout: False

No Addition Operator Overload

While you can find the difference of two sets with -, you cannot use the addition operator + with two sets. This is because there is no clear meaning on what this would do. Do you want the union or the intersection? Because of this ambiguity, it is clearer to use the | and & operators, which also share meaning with standard or and and operations.

Lack Of Order

Because sets are not ordered, you cannot index into them like you can with lists:

1
2

A = { 1, 2, 3 }
print(A[1]) # ERROR: This won't work!

But you can still iterate over them:

1
2
3
4
5
6

A = { 2, 1, 3 }
for element in A:
    print(element)
# stdout:   1   < WOAH, ORDER HAS CHANGED
#           2   <
#           3

Note also how the order is not preserved when you print the elements! DO NOT rely on an ordering when using sets. Even though in the example above the order was “sorted” (i.e. 1 then 2 then 3), you cannot rely on this either.

Implementation Details

CPython implements a set very similar to how it implements a dictionary — with the use of a hash table. The keys are the elements of the set, and the values are “ignored”. Several optimizations can be made due to the lack of values. This hash table implementation allows \( O(1) \) average case membership checking (i.e. is my_element in my_set?).

You can view the source code here.