Question

所以我只是在实践中实现一些经典的排序算法，并提出了用于合并排序的解决方案：

def merge_sort(numbers):
'''Sorts a list based on the traditional merge-sort'''

    if len(numbers) <= 1:
        return numbers

    result = []

    mid = len(numbers) // 2

    left = merge_sort(numbers[mid:])
    right = merge_sort(numbers[:mid])

    l_idx = 0
    r_idx = 0

    # compare all numbers in the two lists
    while l_idx < len(left) and r_idx < len(right):
        if left[l_idx] > right[r_idx]:
            result.append(right[r_idx])
            r_idx += 1
        else:
            result.append(left[l_idx])
            l_idx += 1

    # make sure leftover numbers are added to result
    result += left[l_idx:]
    result += right[r_idx:]

    return result

似乎可行，例如：

>> number_list = [1, 3, 7, 8, 12, 23, 41, 3, 7, 12, 41, 24, 29]
>> sorted(number_list)
[1, 3, 3, 7, 7, 8, 12, 12, 23, 24, 29, 41, 41]
>> merge_sort(number_list)
[1, 3, 3, 7, 7, 8, 12, 12, 23, 24, 29, 41, 41]

输出正确且排序正确，但是当我运行如下所示的单元测试时：

import sorting as s
import unittest
import random
unsorted_list = [random.randint(1, 120) for _ in range(100)]
expected_sorted_list = sorted(unsorted_list)

    def test_mergesort(self):

        merge_sorted_list = s.merge_sort(unsorted_list)
        self.assertEqual(expected_sorted_list, merge_sorted_list)

然后我随后运行：

>> python -m unittest test_sorting
AssertionError: Lists differ: [4, 4[42 chars]2, 13, 14, 15, 15, 16, 17, 18, 19, 20, 21, 22,[308 chars] 120] != [4, 4[42 chars]2, 13]

First list contains 84 additional elements.
First extra element 16:
14

Diff is 762 characters long. Set self.maxDiff to None to see it.

----------------------------------------------------------------------
Ran 2 tests in 0.002s

FAILED (failures=1)

为什么单元测试会引发错误？即使我已经测试了合并排序功能，但似乎排序正确。

目录结构：

/classical-algorithms
    sorting.py
    test_sorting.py

Answer 1

让包含测试的python文件为test_sorting.py。

然后以下对我有用：

import sorting as s
import unittest
import random
unsorted_list = [random.randint(1, 120) for _ in range(100)]
expected_sorted_list = sorted(unsorted_list)

class tests(unittest.TestCase):
    def test_mergesort(self):
        merge_sorted_list = s.merge_sort(unsorted_list)
        self.assertEqual(expected_sorted_list, merge_sorted_list)

运行测试：

$python -m unittest test_sorting
.
----------------------------------------------------------------------
Ran 1 test in 0.000s

OK

Answer 2

我似乎找到了错误。我的旧单元测试文件如下所示：

import sorting as s
import unittest
import random

unsorted_list = [random.randint(1, 120) for _ in range(100)]
expected_sorted_list = sorted(unsorted_list)

    def test_bubblesort(self):

        bubble_sorted_list = s.bubble_sort(unsorted_list)
        self.assertEqual(expected_sorted_list, bubble_sorted_list)

    def test_mergesort(self):

        merge_sorted_list = s.merge_sort(unsorted_list)
        self.assertEqual(expected_sorted_list, merge_sorted_list)

我也在测试我先前实现的bubble_sort。当我将其更改为此：

import sorting as s
import unittest
import random


class TestSorters(unittest.TestCase):

    def test_bubblesort(self):

        unsorted_list = [random.randint(1, 120) for _ in range(100)]
        expected_sorted_list = sorted(unsorted_list)
        bubble_sorted_list = s.bubble_sort(unsorted_list)
        self.assertEqual(expected_sorted_list, bubble_sorted_list)

    def test_mergesort(self):

        unsorted_list = [random.randint(1, 120) for _ in range(100)]
        expected_sorted_list = sorted(unsorted_list)
        merge_sorted_list = s.merge_sort(unsorted_list)
        self.assertEqual(expected_sorted_list, merge_sorted_list)

一切正常。这很奇怪吗？

为什么unittest认为这两个列表不相等？

2 个答案: