为什么 pd.Timedelta modulo int 有效，但不能加上 int？

Question

One can use modulo on Pandas Timedeltas with any number, but addition results in an unsupported error. The following example

import pandas as pd
pd.to_timedelta(7, "s") % 3
pd.to_timedelta(7, "s") + 3

results in

TypeError                                 Traceback (most recent call last)
<ipython-input-2-edd04d4a1971> in <module>
      3 pd.to_timedelta(7, "s") % 3
      4 
----> 5 pd.to_timedelta(7, "s") + 3

TypeError: unsupported operand type(s) for +: 'Timedelta' and 'int'

Is there a deeper reason for this?

Answer 1

you can do it only if you show which part of datetime you want to use, for example:

time=pd.to_timedelta(7, "s")
print(time)
0 days 00:00:07
print(time.seconds%3)
1

You cannot divide using different types

Answer 2

Looking at the code for the Timedelta object , we can see that the __mod__ function just calls the __divmod__ function, and returns the second value; which looks like the following:

def __divmod__(self, other):
    # Naive implementation, room for optimization
    div = self // other
    return div, self - div * other

So % will work with anything that can divide a Timedelta object - we can see what __floordiv__ accepts from the implementation in the same link.