MATLAB冒号(:)运算符的Python等价实现：如何简洁创建单调间隔向量？

Great question! MATLAB's colon (:) operator is one of those super handy shortcuts for creating evenly spaced vectors, and Python has a few equally concise ways to get the same job done. Let's break down the most common use cases with examples that mirror MATLAB's behavior:

1. Basic Integer Vectors (start:end in MATLAB) #

In MATLAB, 1:5 gives you [1, 2, 3, 4, 5]. Since Python's built-in range() is left-closed and right-open, you'll need to adjust the end value by +1 to match:

• Pure Python: Convert the range iterator to a list:

  vec = list(range(1, 6))  # Output: [1, 2, 3, 4, 5]
  

• NumPy (closer to MATLAB's feel): Use np.arange() which returns a numerical array directly:

  import numpy as np
  vec = np.arange(1, 6)  # Output: array([1, 2, 3, 4, 5])
  

2. Vectors with Custom Step Sizes (start:step:end in MATLAB) #

For something like MATLAB's 1:2:5 (which gives [1, 3, 5]), just add the step argument to range() or np.arange():

• Pure Python:

  vec = list(range(1, 6, 2))  # Output: [1, 3, 5]
  # Reverse example (MATLAB's 5:-1:1)
  vec_rev = list(range(5, 0, -1))  # Output: [5, 4, 3, 2, 1]
  

• NumPy:

  vec = np.arange(1, 6, 2)  # Output: array([1, 3, 5])
  vec_rev = np.arange(5, 0, -1)  # Output: array([5, 4, 3, 2, 1])
  

3. Floating-Point Spaced Vectors #

MATLAB makes floating-point steps easy with 0:0.5:2, but Python's range() doesn't support floats. Here's how to handle this:

• Pure Python: Use a list comprehension for simple cases:

  vec = [x * 0.5 for x in range(0, 5)]  # Output: [0.0, 0.5, 1.0, 1.5, 2.0]
  

• NumPy (best for numerical work):
  • Use np.arange() (note: add a tiny buffer to the end value to avoid missing the last element due to floating-point precision):

    vec = np.arange(0, 2.1, 0.5)  # Output: array([0. , 0.5, 1. , 1.5, 2. ])
    

  • Or use np.linspace() when you want to specify the number of points instead of step size (more reliable for floats):

    vec = np.linspace(0, 2, 5)  # Output: array([0. , 0.5, 1. , 1.5, 2. ])
    

Quick Tip #

If you're working in data science, engineering, or any context where you'll be doing numerical operations on the vector, NumPy is your go-to tool—its arrays behave almost identically to MATLAB's vectors. For simple list creation without dependencies, stick to pure Python range() and list comprehensions.

内容的提问来源于stack exchange，提问作者Glenn Jocher