I am struggling with datetime64 comparisons in dataframes to update a column. lets say we have a dataframe 'df' with a 'date' column

df.date.values[0]
Out[128]: numpy.datetime64('2015-05-17T22:00:00.000000000+0800') 

i needed to drop the time and just do a date comparison, so to extract the date i use .date()

df.date[0].date()
Out[131]: datetime.date(2015, 5, 17)

I test it out first

df.date[0].date()==np.datetime64('2015-05-17')
Out[132]: True

So all seems good i try to do a condition if date column equal to a specific date, update another column into another value

df[df.date[0].date()==np.datetime64('2015-05-17')].flag=True

I get a KeyError: True

how do i properly do this?

Have you tried something like:

flag = pd.Series(np.where(df.date == np.datetime64('2015-05-17'), True, False), index=df.index)

flag will be a series of True/False values which you use later.