Sorting by [str, datetime] in order [ascending, descending]

Question

I have an object with two attributes:

i.member : str
i.time : datetime

I tried to sort using:

l.sort(key=lambda i: (i.member, -i.time))

But it tells me that - is invalid with datetime objects (which makes sense). If this were a row in a Pandas data frame, I would sort df.sort_values(by=['member', 'time'], ascending=[True, False]) but I don't want to incur the overhead of loading it just for sorting.

I found previous questions on the topic but they only related to sorting by datetime (but reversed) etc with the reverse parameter, which only takes a boolean (regardless, it is sorting by tuple so I can't pass a list of booleans).

How would I do this in normal Python?

Answer 1

假设您的日期都在 1970 年之后，您可以使用日期的时间戳值进行排序：

l.sort(key=lambda i: (i.member, -i.time.timestamp()))