NumPy

Installation using pip for Python 3

pip3 install numpy

import numpy as np

Convert list to NumPy ndarray

l1 = [1,2,3]

l2 = [4,5,6]

arr1 = np.array(l1)

arr2 = np.array(l2)

print(arr1*arr2)

[ 4 10 18]

NumPy ndarray attributes

• size

• T    #transpose

• ndim

• np.inf #infinite value

print(arr1.ndim)

1

• shape

print(arr1.shape)

(3,)

arr2d = np.array([[1,2,3],

                 [4,5,6]])

print(arr2d.shape)

(2, 3)

NumPy ndarray methods 

• min()

• max()

• mean()

• std()

• sum()

• reshape()

arr2d.reshape(3,2)

array([[1, 2],

       [3, 4],

       [5, 6]])

• flatten()

arr2d.flatten()

array([1, 2, 3, 4, 5, 6])

• tolist()

arr2d.tolist()

[[1, 2, 3], [4, 5, 6]]

NumPy methods

• np.sqrt(value)

• np.zeros(shape)

• np.ones(shape)

• np.full(shape,value)

• np.mean(arr)

• np.median(arr)

• np.log(arr)

• np.where()

  • df['col1']=np.where((df['col2']>=0) & (df['col2']<=100),True,False) #importance: each condition must be surrounded by parentheses

  • df['col1']=np.where(df['col1']==np.inf, 0, df['col1']) #equivalent to the following

  • df.loc[df['col1']==np.inf,'col1']=0