Frequent Python interview questions: shallow copy and deep copy

In Python interviews, shallow copy and deep copy are a relatively difficult question. Sixty percent of people have never heard of shallow copy and deep copy, and thirty percent of people only understand the basics of the two. The difference between concepts and simplicity. Only less than one adult can accurately tell the difference between the two and complete the writing of relevant code demos! Let us solve this difficulty today and add weight for promotion to the interview! This article may also be the most easy-to-understand article on the entire Internet that explains shallow copy and deep copy. As long as you follow the examples in the article and practice, you will definitely understand the technical difficulties of what shallow copy and deep copy are!

Variable assignment

In Python variable data types (lists, dictionaries, sets), assign a variable of a variable data type to another variable. The two variables refer to The same object has the same memory address. If one variable is modified, the other variable will be modified accordingly. If you want to know the technical details of mutable data types and immutable data types in Python, please read the article:​​Frequently asked questions in Python interviews: The difference between mutable data types and immutable data types. ​​

Let’s take an example from my previous article.

l1=['a','b','c']
l2=l1
print(id(l1))
print(id(l2))
l2.append('d')
print("************")
print(id(l1))
print(l1)
print(id(l2))
print(l2)
输出：
838366483528
838366483528
************
838366483528
['a', 'b', 'c', 'd']
838366483528
['a', 'b', 'c', 'd']

The output results will not be explained here. The addresses of l1 and l2 are the same, so they will affect each other.

Shallow copy

Copies the top level of an object and only copies the "reference" of the object.

Take an example from the previous article. There is only one layer in the list, that is, the element does not contain other lists.

l3=['x','y','z']
l4=list(l3)
print(id(l3))
print(id(l4))
l4.append('a')
print(l3)
print(l4)
输出
831456302152
831480344136
['x', 'y', 'z']
['x', 'y', 'z', 'a']

As you can see from the results, the addresses of l3 and l4 are different, so they will not affect each other.

Let’s take another example and let the first element in l3 be a list.

l3=[['x','y','z'],'a','b']
l4=list(l3)
print(id(l3))
print(id(l4))
l4.append('c')
print(l3)
print(l4)
结果
533336249416
533337391240
[['x', 'y', 'z'], 'a', 'b']
[['x', 'y', 'z'], 'a', 'b', 'c']

The result is exactly as expected, because the first element is a list, so it also has an address, and we print it.

print(id(l3[0]))
print(id(l4[0]))
结果
533336248904
533336248904

We found that the addresses of l3[0] and l4[0] are actually the same, and then execute the following code.

l3[0].append('yy')
print(l3)
print(l4)
结果
[['x', 'y', 'z', 'yy'], 'a', 'b']
[['x', 'y', 'z', 'yy'], 'a', 'b', 'c']

The results are in line with expectations, and modifying the value of l3[0] also affects l4[0]. So what can we do to prevent l3[0] and l4[0] from affecting each other? This requires the introduction of deep copies.

Deep copy

Deep copy copies each layer in the object. The copied object is completely independent from the original object and has no relationship. To implement deep copy, you need to use the deepcopy method in the copy module.

import copy
l3=[['x','y','z'],'a','b']
#l4=list(l3) #浅拷贝，使用=copy.copy(l3)也可以实现浅拷贝
l4=copy.deepcopy(l3) #深拷贝
l4.append('c')
print(id(l3[0]))
print(id(l4[0]))
l3[0].append('yy')
print(l3)
print(l4)
输出
407168435784
407166887304
[['x', 'y', 'z', 'yy'], 'a', 'b']
[['x', 'y', 'z'], 'a', 'b', 'c']

You can see that l3[0] and l4[0] are completely independent and have no influence on each other.

Summary of shallow copy and deep copy

Not much to say, just use two pictures to give a simple explanation:

Shallow copy, as shown below:

Deep copy, as shown below:

The above is the detailed content of Frequent Python interview questions: shallow copy and deep copy. For more information, please follow other related articles on the PHP Chinese website!