pandas questions about argmin and timestamp - pandas

Question

final_month = pd.Timestamp('2018-02-01')
df_final_month = df[df['week'] >= final_month]
df_final_month.iloc[:, 1:].sum().argmax()
index = df.set_index('week')
index['storeC'].argmin()
the code above is correct, i just don't exactly understand how does it work inside. i have some questions:
1.the type(week) is datetime, the reason why set final_month as Timestamp is that the datetime is almost as same as Timestamp, they recognise each other in Python?
2.about the argmax(), and argmin(), for the df_final_month.iloc[:, 1:].sum().argmax(), i removed sum() and tried like df_final_month.iloc[:, 1:].argmax(), it returns
`AttributeError: 'DataFrame' object has no attribute 'argmax'`
why is it? why the second code doesn't need a max() or something to call argmin(), what's the requirement for using argmin()/argmax() ?
please explaining the details of how python or pandas deal with these data, the more detail the better.
thanks!
i am new in Python.

Answer 1

Is Timestamp almost as same as datetime?
Here is quote from pandas documentation itself:
TimeStamp is the pandas equivalent of python’s Datetime and is interchangable with it in most cases
In fact, if you look at source code of pandas you will see that Timestamp actually inherits from datetime. Here is code to check these statements are true:
dt = datetime.datetime(2018, 1, 1)
ts = pd.Timestamp('2018-01-01')
dt == ts # True
isinstance(ts, datetime.datetime) # True
Why calling argmax method on DataFrame, without calling sum throws an error?
Because DataFrame object doesn't have argmax method, only Series do. And sum, in your case, returns a Series instance.