dpnp.ndarray.view

method

ndarray.view(dtype=None, *, type=None)

New view of array with the same data.

For full documentation refer to numpy.ndarray.view.

Parameters:

• dtype ({None, str, dtype object, type}, optional) --

  The desired data type of the returned view, e.g. dpnp.float32 or dpnp.int16. Omitting it results in the view having the same data type. Can also be a subclass of dpnp.ndarray to create a view of that type (this is equivalent to setting the type parameter).

  Default: None.

• type ({None, type}, optional) --

  Type of the returned view, e.g. a subclass of dpnp.ndarray. If specified, the returned array will be an instance of type. Omitting it results in type preservation.

  Default: None.

Notes

Passing None for dtype is the same as omitting the parameter, opposite to NumPy where they have different meaning.

view(some_dtype) or view(dtype=some_dtype) constructs a view of the array's memory with a different data type. This can cause a reinterpretation of the bytes of memory.

Only the last axis has to be contiguous.

Examples

>>> import dpnp as np
>>> x = np.ones((4,), dtype=np.float32)
>>> xv = x.view(dtype=np.int32)
>>> xv[:] = 0
>>> xv
array([0, 0, 0, 0], dtype=int32)

However, views that change dtype are totally fine for arrays with a contiguous last axis, even if the rest of the axes are not C-contiguous:

>>> x = np.arange(2 * 3 * 4, dtype=np.int8).reshape(2, 3, 4)
>>> x.transpose(1, 0, 2).view(np.int16)
array([[[ 256,  770],
        [3340, 3854]],

    [[1284, 1798],
        [4368, 4882]],

    [[2312, 2826],
        [5396, 5910]]], dtype=int16)

Creating a view with a custom ndarray subclass:

>>> class MyArray(np.ndarray):
...     pass
>>> x = np.array([1, 2, 3])
>>> y = x.view(MyArray)
>>> type(y)
<class 'MyArray'>