TypeError: concatenate() 得到了一个意外的关键字参数“dtype”

Question

I can clearly see an argument named dtype in numpy documentation .

What I can't do, is this:

np.concatenate((np.array([]),np.array([3,4])),dtype=np.int64)

Which gives this error:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<__array_function__ internals>", line 4, in concatenate
TypeError: concatenate() got an unexpected keyword argument 'dtype'

Without the dtype argument I get this array:

>>> np.concatenate((np.array([]),np.array([3,4])))
array([3., 4.])

But I don't want those to be decimal points (float) values.

Why can't I use the dtype argument? How to make it int on concatenation even if my initial arrays had let's say float values?

Answer 1

The documentation for dtype says:

New in version 1.20.0.

Version 1.20.0 has only been released in Jan 2021 .

Check if you have the right version installed.

Answer 2

If you look at the documentation of version 1.19 , dtype was not an argument of concatenate . So dtype was introduced in version 1.20.

Then you can update your numpy version. Otherwise, if your first array is of type np.int64 and the second one contains only int , the resulting type is np.int64 .

np.concatenate((np.array([], dtype=np.int64), np.array([3, 4])))

array([3, 4])