如何在数据帧中向TimedeltaIndex添加偏移量

Question

如何使用TimedeltaIndex向Panda数据帧添加偏移量

dfN.index

产出：

TimedeltaIndex([       '00:00:00', '00:00:00.100000', '00:00:00.200000',
                '00:00:00.300000', '00:00:00.400000', '00:00:00.500000',
                '00:00:00.600000', '00:00:00.700000', '00:00:00.800000',
                '00:00:00.900000',
                ...
                '01:56:26.700000', '01:56:26.800000', '01:56:26.900000',
                       '01:56:27', '01:56:27.100000', '01:56:27.200000',
                '01:56:27.300000', '01:56:27.400000', '01:56:27.500000',
                '01:56:27.600000'],
               dtype='timedelta64[ns]', name='timestamps', length=69877, freq='100L')

这不起作用：

dfN.index += Offset #Offset = 100

输出

TypeError: Addition/subtraction of integers and integer-arrays with TimedeltaArray is no longer supported.  Instead of adding/subtracting `n`, use `n * obj.freq`

在这种情况下，数据间隔为每100毫秒。我想加上和抵消500毫秒

Answer 1

变量Offset不能是整数。您需要创建一个Timedelta对象来添加它。试试这个：

dfN.index += pd.Timedelta(100, unit=‘milliseconds’) #This will add 100 miliseconds

或

dfN.index += pd.Timedelta(milliseconds = 100)