我通过将datetime导入为realdatetime并将我在模拟中需要的方法替换为实际方法来完成这项工作:

import datetime as realdatetime

@mock.patch('datetime')
def test_method(self, mock_datetime):
    mock_datetime.today = realdatetime.today
    mock_datetime.now.return_value = realdatetime.datetime(2019, 8, 23, 14, 34, 8, 0)

几天前我遇到了同样的情况，我的解决方案是在模块中定义一个函数来测试和模拟:

def get_date_now():
    return datetime.datetime.now()

今天我发现了《FreezeGun》，它似乎很好地解决了这个问题

from freezegun import freeze_time
import datetime
import unittest


@freeze_time("2012-01-14")
def test():
    assert datetime.datetime.now() == datetime.datetime(2012, 1, 14)

也许您可以使用自己的“today()”方法，在需要的地方打补丁。使用utcnow()的示例可以在这里找到:https://bitbucket.org/k_bx/blog/src/tip/source/en_posts/2012-07-13-double-call-hack.rst?at=default

对于那些在测试类中使用补丁程序的人，下面是我如何成功地修补datetime功能:

from datetime import datetime
import unittest
from unittest.mock import Mock, patch

# Replace with the proper path to the module you would
# like datetime to be mocked
from path.to.my_module

class MyTestCases(unittest.TestCase):

    def setUp(self):
        """execute on class instantiation"""
        # Record both times at the same moment
        self.dt_now, self.dt_utcnow = datetime.now(), datetime.utcnow()

        # After retrieving real (or hardcoded datetime values), 
        # proceed to mock them in desired module
        self.patch_datetime_functions()


    def patch_datetime_functions(self) -> None:
        """
        Patch datetime.now() and datetime.utcnow() to prevent issues when
        comparing expected dates
        """

        # Create a patcher
        self.patcher_dt = patch(
            'path.to.my_module'
        )

        # Start but make sure cleanup always occurs
        self.patcher_dt.start()
        self.addCleanup(self.patcher_dt.stop)

        # Perform the actual patch – use lambdas as mock functions
        datetime_mock = Mock(wraps=datetime)
        datetime_mock.now.return_value = self.dt_now
        datetime_mock.utcnow.return_value = self.dt_utcnow

        my_module.datetime = datetime_mock


    # Here's what it will look like when testing:
    def some_test(self):
        curr_dt = self.dt_now
        returned_dt = my_module.datetime.utcnow()
        
        # Compare the dates
        self.assertEqual(curr_dt, returned_dt,
            'Datetime values should be equal'
        )

对我来说，最好的方法是结合@Daniel G和@frx08的解决方案:

class Test_mock_date:
    class NewDate(datetime.datetime):
        @classmethod
        def now(cls, tz=None):
            return cls(2021, 5, 12)

    def test_mock_date(self):
        with patch('datetime.datetime', new = self.NewDate):
            assert datetime.datetime.now() == datetime.datetime(2021, 5, 12, 0, 0)

你可以看看下面我写的关于如何使用MagicMock的不同示例的中型文章https://medium.com/@camposer/d2113513b365

CPython实际上使用纯python Lib/datetime.py和c优化的Modules/ _datetimmodule .c实现了datetime模块。c优化版本不能打补丁，但纯python版本可以。

在Lib/datetime.py中纯python实现的底部是这样的代码:

try:
    from _datetime import *  # <-- Import from C-optimized module.
except ImportError:
    pass

这段代码导入了所有c优化的定义，并有效地替换了所有纯python定义。我们可以通过以下方式强制CPython使用datetime模块的纯python实现:

import datetime
import importlib
import sys

sys.modules["_datetime"] = None
importlib.reload(datetime)

通过设置sys。modules["_datetime"] = None，我们告诉Python忽略c优化的模块。然后重新加载导致从_datetime导入失败的模块。现在纯python定义仍然存在，可以正常打补丁。

如果你正在使用Pytest，那么在conftest.py中包含上面的代码片段，你就可以正常地修补datetime对象了。