如果没有 NumPy，我应该如何将两个向量或矩阵相减？

Question

I have a function that subtracts one matrix of some shape (a nested list) from another matrix of the same shape (assumed to be a nested list of the same size):

def subtract(matrix_a, matrix_b):
    MATRIX_LENGTH = len(matrix_a)
    MATRIX_ROW = len(matrix_a[0])
    for i in range(MATRIX_LENGTH):
      for j in range(MATRIX_ROW):
        matrix_a[i][j] -= matrix_b[i][j]
    return matrix_a

The above function iterates through each value of matrix_a and subtracts the value at the same position of matrix_b from matrix_a before returning matrix_a.

This code does not work if I supply a vector (an "un"-nested list). For example, supplying [1, 2, 3, 4] and [2, 3, 4, 5] to matrix_a and matrix_b respectively will yield TypeError: object of type 'int' has no len() .

How should I modify the function to accept "un"-nested lists? For educational purposes, I do not want to use NumPy.

Thanks in advance.

Answer 1

Thanks to @LarrytheLlama, I thought of using a try/except/else statement to catch the TypeError. The following code subtracts two vectors or matrices from each other without NumPy:

def subtract(matrix_a, matrix_b):
  MATRIX_LENGTH = len(matrix_a)
  try:
    MATRIX_ROW = len(matrix_a[0])
  except TypeError:
    for i in range(MATRIX_LENGTH):
      matrix_a[i] -= matrix_b[i]
  else:
    for i in range(MATRIX_LENGTH):
      for j in range(MATRIX_ROW):
        matrix_a[i][j] -= matrix_b[i][j]
  return matrix_a

Answer 2

You can use zip to pair up the values/rows in a list comprehension, and call the function recursively for nested dimensions:

def subtract(A,B):
     if isinstance(A,list):
         return [ subtract(ra,rb) for ra,rb in zip(A,B) ]
     else:
         return A-B

This will work for any number of dimensions