Data Science Essentials in Python

Unit 21

Creating Arrays

numpy arrays are more compact and faster than native Python lists, especially in multidimensional cases. However, unlike lists, arrays are homogeneous: you cannot mix and match array items that belong to different data types.

There are several ways to create a numpy array. The function array creates an array from array-like data. The data can be a list, a tuple, or another array. numpy infers the type of the array elements from the data, unless you explicitly pass the dtype parameter. numpy supports close to twenty data types, such as bool_, int64, uint64, float64, and <U32 (for Unicode strings).

When numpy creates an array, it doesn’t copy the data from the source to the new array, but it links to it for efficiency ...