# numpy Arrays Numpy n-dimensional array: the ndarray

## Example

The core data structure in numpy is the `ndarray` (short for n-dimensional array). `ndarray`s are

• homogeneous (i.e. they contain items of the same data-type)
• contain items of fixed sizes (given by a shape, a tuple of n positive integers that specify the sizes of each dimension)

One-dimensional array:

``````x = np.arange(15)
# array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14])
x.shape
# (15,)
``````

Two-dimensional array:

``````x = np.asarray([[0, 1, 2, 3, 4], [5, 6, 7, 8, 9], [10, 11, 12, 13, 14]])
x
# array([[ 0, 1, 2, 3, 4],
# [ 5, 6, 7, 8, 9],
# [10, 11, 12, 13, 14]])
x.shape
# (3, 5)
``````

Three-dimensional:

``````np.arange(12).reshape([2,3,2])
``````

To initialize an array without specifying its contents use:

``````x = np.empty([2, 2])
# array([[ 0., 0.],
# [ 0., 0.]])
``````

Datatype guessing and automatic casting

The data-type is set to float by default

``````x = np.empty([2, 2])
# array([[ 0., 0.],
# [ 0., 0.]])

x.dtype
# dtype('float64')
``````

If some data is provided, numpy will guess the data-type:

``````x = np.asarray([[1, 2], [3, 4]])
x.dtype
# dtype('int32')
``````

Note that when doing assignments numpy will attempt to automatically cast values to suit the `ndarray`'s datatype

``````x[1, 1] = 1.5 # assign a float value
x[1, 1]
# 1
# value has been casted to int
x[1, 1] = 'z' # value cannot be casted, resulting in a ValueError
``````

``````x = np.asarray([[1, 2], [3, 4]])
# array([[1, 2],
[3, 4]])
y = np.asarray([[5, 6]])
# array([[5, 6]])
``````

In matrix terminology, we would have a 2x2 matrix and a 1x2 row vector. Still we're able to do a sum

``````# x + y
array([[ 6, 8],
[ 8, 10]])
``````

This is because the array `y` is "stretched" to:

``````array([[5, 6],
[5, 6]])
``````

to suit the shape of `x`.

Resources: PDF - Download numpy for free