python之使用这种日期格式对列表进行反向排序的最佳方法是什么

我有一个包含数百万条记录的数据库:

[("Something", "10/08/2017", "something else", "something more", "etc"), 
("Something", "16/08/2017", "something else", "something more", "etc"), 
("Something", "14/07/2017", "something else", "something more", "etc"), 
("Something", "14/08/2017", "something else", "something more", "etc"), 
("Something", "15/07/2017", "something else", "something more", "etc"), 
("Something", "14/08/2017", "something else", "something more", "etc")] 

那是 DD/MM/YYYY 我需要它按第二个值反向排序。

第一次尝试:
我试过了，但它只按天排序，而不是按月排序。

def sort_by(my_list, index): 
    my_list.sort(key=lambda x: x[index], reverse=True) 
    return my_list 

结果图片(忽略颜色，它来 self 的 GUI):https://image.prntscr.com/image/ZeIgzxNaQ2OD4VX0Yztnhw.png
如您所见，它只按天排序，而不是按月排序，因此每次日期变化时，它都会从一个月跳到另一个月。真是一团糟。

第二次尝试:

def sort_by_date(my_list, index): 
    for record, data in enumerate(my_list): 
        split = data[index].split("/") 
        altered_record = my_list[record] 
        altered_record[index] = split[2]+"-"+split[1]+"-"+split[0] 
    my_list.sort(key=lambda x: x[index], reverse=True) 
    return my_list 

那时我意识到它不会工作，因为元组不允许项目分配，所以我必须将整个数据转换为列表列表而不是元组列表。即使我这样做了，我也怀疑它是否有效，而且它会花费很多时间来完成这项工作。

所以我的问题是:
您将如何按 DD/MM/YYYY 对元组列表进行排序？

根据@juanpa.arrivillaga 的评论，所需的输出是按日期反向排序的列表:

31/12/2017 
30/12/2017 
29/12/2017 
28/12/2017 
(And the next days of that month) 
30/11/2017 
29/11/2017 
28/11/2017 
27/11/2017 
(And so go on) 

编辑:如果可能，我想将日期保留为字符串，而不更改日期类型。

请您参考如下方法:

您可以为此使用 Python 的 datetime 模块:

from pprint import pprint 
from datetime import datetime 
 
database = [("Something", "10/08/2017", "something else", "something more", "etc"), 
            ("Something", "16/08/2017", "something else", "something more", "etc"), 
            ("Something", "14/07/2017", "something else", "something more", "etc"), 
            ("Something", "14/08/2017", "something else", "something more", "etc"), 
            ("Something", "15/07/2017", "something else", "something more", "etc"), 
            ("Something", "14/08/2017", "something else", "something more", "etc")] 
 
pprint(sorted(database, key=lambda x: datetime.strptime(x[1], "%d/%m/%Y"), reverse=True)) 

有输出:

[('Something', '16/08/2017', 'something else', 'something more', 'etc'), 
 ('Something', '14/08/2017', 'something else', 'something more', 'etc'), 
 ('Something', '14/08/2017', 'something else', 'something more', 'etc'), 
 ('Something', '10/08/2017', 'something else', 'something more', 'etc'), 
 ('Something', '15/07/2017', 'something else', 'something more', 'etc'), 
 ('Something', '14/07/2017', 'something else', 'something more', 'etc')]