将Numpy数组转换为datetime.date pandas DataFrame

Question

I have a numpy.ndarray thats values are pandas._libs.tslib.Timestamp

Example:

In: type(closedDate)

Out: numpy.ndarray

In: type(closedDate[0])

Out: pandas._libs.tslib.Timestamp

I would like to convert the contents of closedDate into a list of datetime.date

I have tried the following:

for i in closedDate:
    closedDate[i].to_datetime()

But get this error:

IndexError: only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices

What to do? :/

Answer 1

I'd expect numpy to provide this functionality via astype , so you would not even have to loop over the array explicitely:

dt_arr = closedDate.astype(np.datetime64)

check:

closedDate
Out: 
array([Timestamp('2017-09-25 14:39:00'), Timestamp('2017-09-26 14:39:00'),
   Timestamp('2017-09-27 14:39:00')], dtype=object)

type(closedDate)
Out: numpy.ndarray

type(closedDate[0])
Out: pandas._libs.tslib.Timestamp

type(dt_arr)
Out: numpy.ndarray

type(dt_arr[0])
Out: numpy.datetime64

Answer 2

[x.to_pydatetime().date() for x in closedDate]

Note, however, that you've probably made some kind of mistake to end up with a numpy array containing a pandas datatype. pandas Series and DataFrames are better equipped to handle pandas-specific types than numpy arrays are.