Why does numpy.datetime64 give an error with an array of years?

user_271526 0
Viewed 690

I am working with some data that has the date in the format 2017.08333. In the process of converting this to datetime objects I got an error when trying to convert the integer years to datetime. 

np.datetime64(['2017','2018'])

The above fails with the error:Could not convert object to NumPy datetime

I was under the impression numpy functions were vectorized. Can someone explain why this doesn't work and how I can solve this? I have 3,000 dates to convert so something that doesn't involve a loop is ideal.

numpy datetime python
3 Answers

To create an array of dates, you could do:

import numpy as np

print(np.array(['2018', '2017'], dtype=np.datetime64))

Output

['2018' '2017']

A more detailed explanation can be found here

user_40015920

Some of them are. I'm not sure what your input looks like (list or np.array), but you could do something like

np.array(['2017', '2018'], dtype=np.datetime64)

or 

my_np_array.astype(np.datetime64)
user_16103880

You can use map as additional too:

>>> import numpy as np
>>> np.array(list(map(np.datetime64,['2017', '2018'])))
array(['2017', '2018'], dtype='datetime64[Y]')
user_87083640
Related