In this lesson, you will learn why copying nested structures in Python requires extra caution to avoid accidentally modifying your original data.
The Hidden Trap of Shallow Copies
When you create a "shallow copy" of a list—using methods like .copy(), the list() constructor, or slicing [:]—Python creates a new container for the outer list. However, it does not create new copies of the objects inside that list. Instead, it simply copies the memory references to those inner objects.
If those inner objects are mutable, such as nested lists, both your original list and your new copy will point to the exact same inner lists in memory. If you modify an inner list through one, the change will appear in both.
Illustrating the Risk
Consider this scenario where we try to copy a list containing a sub-list:
original = [[1, 2], [3, 4]]
shallow_copy = original.copy()
# Modify the nested list through the copy
shallow_copy[0][0] = 99
print(original) # Output: [[99, 2], [3, 4]]
print(shallow_copy) # Output: [[99, 2], [3, 4]]
Even though shallow_copy is a different object than original, they both share the same inner list object at index 0. When we change the first element of the first inner list, the "original" data is corrupted. This is a common source of bugs when developers assume a copy is completely independent of the source.
Handling Nested Sequences
To prevent this, you must distinguish between the outer structure and its contents. If you only have one level of nesting and need a completely independent copy, you would typically use the copy module’s deepcopy() function. Deep copying recursively traverses the structure, creating new copies of every nested object it finds. By being mindful of whether your data contains mutable "containers within containers," you can choose the right copying strategy to preserve data integrity and prevent unintended side effects.
Key Takeaway: A shallow copy only duplicates the outer container; nested mutable objects remain shared references, meaning changes to them will affect both the copy and the original.