如何在datetime上对类进行排序,对collections.deque进行排序

时间:2013-11-05 18:00:37

标签: python unit-testing sorting python-2.7 comparison

我真的可以使用一些帮助来实现 getitem iter 方法或生成器函数来对我创建的类和类容器进行排序。

我创建了一个具有send_time(datetime)和period_length(int)属性的Report类。 我还为Reports创建了一个ReportDeque容器,它继承自collections.deque。

我需要为类和它的容器添加排序功能。

到目前为止,我已经将工作分类了,但是想让list.sort()样式正常工作。

  sortedList = sorted(list, key=lambda report: report.send_time)
  sortedDeque = sorted(deque, key=lambda report: report.send_time)

我正在努力在Report中实现 getitem ,以及 iter ,在ReportDeque中实现下一个方法。我似乎无法找到我需要的所有这些工作的例子。

也许应该使用生成器函数来对collections.deque容器进行排序。 有各种各样的生成器以不同的方式对deque进行排序会很好。

以下是我的测试用例。要在下面的代码中运行unittest,请输入:

  python -m unittest test_reports

输出结束于此帖。

提前致谢...

------------------- test_reports.py snip ------------------------ -

#!/usr/bin/env python

from datetime import datetime
from collections import deque
import unittest
import inspect

class Report(object):
    """
    Contains all information contained in a report
    """
    def __init__(self, periodStart, periodEnd, sendTime):
        self.period_start = periodStart
        self.period_end = periodEnd
        self.send_time = sendTime
        self.send_timestamp = (sendTime - datetime(1970, 1, 1)).total_seconds()
        self.period_length = (periodEnd - periodStart).total_seconds()

    #def __getitem__(self, key):

class ReportDeque(deque):
    """
    Container for processing, sorting Report objects
    """

    #def __iter__(self)

    #def next(self)

class TestReports(unittest.TestCase):

    def setUp(self):

        self.list = []
        self.deque = ReportDeque()

        # send_time 12/4/13, day length report
        report = Report(datetime(2013, 12, 3, 0), datetime(2013, 12, 3, 23), datetime(2013, 12, 4, 0))
        self.list.append(report)
        self.deque.append(report)
        # send_time 12/3/13, day length report
        report = Report(datetime(2013, 12, 2, 0), datetime(2013, 12, 2, 23), datetime(2013, 12, 3, 0))
        self.list.append(report)
        self.deque.append(report)
        # send_time 12/2/13, day length report
        report = Report(datetime(2013, 12, 1, 0), datetime(2013, 12, 1, 23), datetime(2013, 12, 2, 0))
        self.list.append(report)
        self.deque.append(report)

        # sorted with key function works
        self.sortedList = sorted(self.list, key=lambda report: report.send_time)
        self.sortedDeque = sorted(self.deque, key=lambda report: report.send_time)

    def test_sort_deque_send_time(self):
        self.print_inspect()
        # deque does not have sort method. How to sort it?
        self.deque.sort()
        firstReport = self.deque[0]
        print "send_time {} period_length {}".format(firstReport.send_time, firstReport.period_length)
        self.assertEqual(firstReport.send_time, datetime(2013, 12, 2, 0, 0, 0, 0))

    def test_sort_list_send_time(self):
        self.print_inspect()
        # list.sort() not working. How to implement __get_item___?
        self.list.sort()
        firstReport = self.list[0]
        print "send_time {} period_length {}".format(firstReport.send_time, firstReport.period_length)
        self.assertEqual(firstReport.send_time, datetime(2013, 12, 2, 0, 0, 0, 0))

    def test_sorted_deque_send_time(self):
        self.print_inspect()
        firstReport = self.sortedDeque[0]
        print "send_time {} period_length {}".format(firstReport.send_time, firstReport.period_length)
        self.assertEqual(firstReport.send_time, datetime(2013, 12, 2, 0, 0, 0, 0))

    def test_sorted_list_send_time(self):
        self.print_inspect()
        firstReport = self.sortedList[0]
        print "send_time {} period_length {}".format(firstReport.send_time, firstReport.period_length)
        self.assertEqual(firstReport.send_time, datetime(2013, 12, 2, 0, 0, 0, 0))

    def print_inspect(self):
        calling_function = inspect.stack()[1][3]
        print "\nin {}()".format(calling_function)


if __name__ == "__main__":
    unittest.main()

------------------- test_reports.py snip ------------------------ -

    $ python -m unittest test_reports


    in test_sort_deque_send_time()
    E
    in test_sort_list_send_time()
    send_time 2013-12-04 00:00:00 period_length 82800.0
    F
    in test_sorted_deque_send_time()
    send_time 2013-12-02 00:00:00 period_length 82800.0
    .
    in test_sorted_list_send_time()
    send_time 2013-12-02 00:00:00 period_length 82800.0
    .
    ======================================================================
    ERROR: test_sort_deque_send_time (test_reports.TestReports)
    ----------------------------------------------------------------------
    Traceback (most recent call last):
      File "test_reports.py", line 51, in test_sort_deque_send_time
        self.deque.sort()
    AttributeError: 'ReportsDeque' object has no attribute 'sort'

    ======================================================================
    FAIL: test_sort_list_send_time (test_reports.TestReports)
    ----------------------------------------------------------------------
    Traceback (most recent call last):
      File "test_reports.py", line 62, in test_sort_list_send_time
        self.assertEqual(firstReport.send_time, datetime(2013, 12, 2, 0, 0, 0, 0))
    AssertionError: datetime.datetime(2013, 12, 4, 0, 0) != datetime.datetime(2013, 12, 2, 0, 0)

----------------------------------------------------------------------
Ran 4 tests in 0.011s

FAILED (failures=1, errors=1)

1 个答案:

答案 0 :(得分:2)

首先,您需要使报表对象具有可比性,因此您可以在没有明确键的情况下进行订购。您可能应该阅读丰富的比较,但__cmp__可以解决这个问题。

class Report(object):
    """
    Contains all information contained in a report
    """
    def __init__(self, periodStart, periodEnd, sendTime):
        self.period_start = periodStart
        self.period_end = periodEnd
        self.send_time = sendTime
        self.send_timestamp = (sendTime - datetime(1970, 1, 1)).total_seconds()
        self.period_length = (periodEnd - periodStart).total_seconds()

    def __cmp__(self, other):
        return cmp(self.send_time, other.send_time)

在测试list.sort()时,这就是让测试通过的全部内容。 sorted(list)sorted(deque)的测试也应该有效,但有一个问题。由于您正在寻求__getitem__实施方面的帮助,我认为您认为sorted()正在进行此类排序,并会对您的双端队列进行排序。这不是它的工作原理。 sorted(iterable)将返回一个新的排序列表,其中包含您的iterable项。

如果你真的想要对你的双端队列进行排序,你必须在deque.sort()方法中实现deque排序算法来做到这一点,我不知道哪种算法更有效率排序deque(我甚至不确定这样做是否有意义),但我认为你可能更容易重建deque并利用python非常有效的排序算法:

class ReportDeque(deque):
    """
    Container for processing, sorting Report objects
    """

    def sort(self, *args, **kwargs):        
        items = [self.pop() for x in xrange(len(self))]
        items.sort(*args, **kwargs)
        self.extend(items)

这应该让你的所有考试都通过。

<强>更新

如果您希望在send_time相等时使用period_length消除歧义,则只需将其添加到__cmp__,如下所示:

    def __cmp__(self, other):
         cmp((self.send_time, self.period_length), 
             (other.send_time, other.period_length))