Understanding the Strange Behaviour of Arrays in Python with NumPy

Dive into the reasons behind unexpected results when manipulating arrays in Python using NumPy and learn how to properly create arrays to avoid confusion.
---

Visit these links for original content and any more details, such as alternate solutions, latest updates/developments on topic, comments, revision history etc. For example, the original title of the Question was: Strange comportement of list/array

If anything seems off to you, please feel free to write me at vlogize [AT] gmail [DOT] com.
---
Understanding the Strange Behaviour of Arrays in Python with NumPy

When working with arrays in Python, particularly with libraries like NumPy, encountering odd behaviours can be frustrating. A common issue is related to how arrays are initialized and manipulated. In this post, we'll explore a specific problem that arises when trying to assign values to a NumPy array, accompanied by the implications of using Python's list multiplication for array creation.

The Problem

You may find yourself in a situation like this:

[[See Video to Reveal this Text or Code Snippet]]

After running this code, instead of receiving an array filled with the expected calculated values, you may notice that P still contains mostly zeros, except for a single one at position (2, 2). This can be perplexing, especially since the printed values reflect calculations that should modify the array.

Understanding the Behaviour

The root of this issue lies in how the array is constructed. When you create an array in this manner ([[0]*n]*n), you aren't actually creating a two-dimensional array but rather creating references to the same list in multiple locations. Here’s a breakdown of the problem:

Mutable vs Immutable: In Python, lists are mutable, meaning that if you modify one instance of a reference, all other references to it will reflect that change. Therefore, all rows in P point to the same row, and modifying one part of that row will affect all others.

Array Data Type: By default, NumPy infers the data type from the first value in the array. When initializing with integers, if you later try to assign a floating-point value, NumPy will clip this value to fit the integer data type, leading to loss of information.

Demonstrating the Issue

You can observe the data type behaviour using the following example:

[[See Video to Reveal this Text or Code Snippet]]

When you attempt to assign 0.3 to an integer array, it gets truncated to 0 since the array type is int.

The Solution

To get the functionality you expect, you have two primary solutions to create a proper two-dimensional array in NumPy:

Option 1: Initialize with Float Values

You can create the array using floats from the outset:

[[See Video to Reveal this Text or Code Snippet]]

Option 2: Specify Data Type Explicitly

Alternatively, you can specify the desired data type directly when creating the array:

[[See Video to Reveal this Text or Code Snippet]]

Resulting Array

With either method, the resulting array P will behave as expected, properly retaining all floating-point calculations:

[[See Video to Reveal this Text or Code Snippet]]

Conclusion

Understanding how to properly set up and manipulate arrays in NumPy is crucial for when you are performing computations. This insight into mutable references and data types can save you from a lot of confusion. Always ensure you're initializing your arrays appropriately to get the intended results. Happy coding!