如何修正生成{0,1}与二元组配对全组合的代码错误？

Hey there! Let's fix this up and get you the exact 16-element list you're looking for. First, let's clarify the goal: we need every possible way to assign a 0 or 1 to each of the four 2D tuples (0,0), (0,1), (1,0), (1,1). Each assignment will be a list of 4 ((x,y), z) entries, and since there are 2 choices per tuple, we get 2⁴=16 total combinations.

Why Your Previous Approaches Didn't Work #

• numpy.meshgrid mistake: You tried combining the tuple set with {0,1} directly, but meshgrid wasn't structured to map each tuple to a unique 0/1 value across all combinations. The output you got was a messy multi-dimensional array instead of the paired assignments you needed.
• itertools.combinations is the wrong tool: This function generates subsets of elements (e.g., pairs of tuples), but we don't want subsets—we want to assign a 0/1 to every tuple in the set, covering all possible assignment permutations.

The Correct Solution with itertools.product #

itertools.product is perfect here because it generates the Cartesian product of sets, which is exactly what we need for all 0/1 assignment combinations. Here's the step-by-step code:

import itertools

# Step 1: Create the set of 2D binary tuples
bset = [0, 1]
fset = list(itertools.product(bset, bset))  # Gives [(0,0), (0,1), (1,0), (1,1)]

# Step 2: Generate all possible 0/1 sequences for 4 elements (one per tuple)
# `repeat=4` means we're creating sequences of length 4, each element 0 or 1
z_assignments = itertools.product(bset, repeat=4)

# Step 3: Pair each tuple with its corresponding 0/1 value in each sequence
final_result = [list(zip(fset, z_seq)) for z_seq in z_assignments]

Verify the Result #

• len(final_result) will be 16 (exactly the number of combinations we need)
• Each inner list has 4 entries, e.g.:
  • First entry: [((0, 0), 0), ((0, 1), 0), ((1, 0), 0), ((1, 1), 0)] (all zeros assigned)
  • Middle entry example: [((0, 0), 0), ((0, 1), 1), ((1, 0), 0), ((1, 1), 1)]
  • Last entry: [((0, 0), 1), ((0, 1), 1), ((1, 0), 1), ((1, 1), 1)] (all ones assigned)

Alternative NumPy Approach #

If you prefer using NumPy, you can adjust your meshgrid usage to generate the 0/1 sequences correctly:

import numpy as np

bset = np.array([0, 1])
# Create the 2D tuple set (same as before)
fset = np.array(np.meshgrid(bset, bset)).T.reshape(-1, 2)
fset = [tuple(i) for i in fset]

# Generate all 4-length 0/1 sequences using meshgrid
z_assignments = np.array(np.meshgrid(bset, bset, bset, bset)).T.reshape(-1, 4)

# Pair tuples with assignments
final_result = [list(zip(fset, z_seq)) for z_seq in z_assignments]

This will give you the exact same 16-element list as the itertools version.

内容的提问来源于stack exchange，提问作者user11562913