Find max value & its index in Numpy Array | numpy.amax()

/ Python / By

Finding max value and it’s index in Numpy Array

In this article we will discuss about how we can find max value and the index of that max value in Numpy array using numpy.amx().

numpy.amax( ) :

Syntax-numpy.amax(arr, axis=None, out=None, keepdims=<no value>, initial=<no value>)

Parameters :

  1. arr: Numpy array
  2. axis: This is an optional parameter unless provided flattens the array.

(Default: Returns the array of max values)

  1. Axis = 0: Returns array containing max values of each columns
  2. Axis = 1: Returns array containing max values of each rows

Let’s see one by one how to find it in 1D and 2D Numpy array.

Maximum value & its index in a 1D Numpy Array:

Let’s create a 1D numpy array from a list given below and find the maximum values and its index

Find maximum value:

To find the maximum value in the array, we can use numpy.amax( ) function and pass the array as function to it.

[10,5,19,56,87,96,74,15,50,12,98]
import numpy as np

# Finding the maximum value inside an array using amax( )
arr = np.array([10, 5, 19, 56, 87, 96, 74, 15, 50, 12, 98])
maxElem = np.amax(arr)
print("Max element : ", maxElem)

Output :
Max element :  98

Find index of maximum value :

To get the index of the max value in the array, we will we have to use the where( ) function from the numpy library.

CODE:

import numpy as np

# Index of the maximum element
arr = np.array([10, 5, 19, 56, 87, 96, 74, 15, 50, 12, 98])
maxElem = np.amax(arr)
print("Max element : ", maxElem)
res = np.where(arr == np.amax(arr))
print("Returned result  :", res)
print("List of Indices of maximum element :", res[0])

Output :
Max element :  98
Returned result  : (array([10], dtype=int32),)
List of Indices of maximum element : [10]

Here, when we called the where( ) function, it returned us with a tuple of  arrays containing the indices of all the  values that follow our conditions.

Find maximum value & its index in a 2D Numpy Array

We will use the following 2D array to demonstrate.

{50,59,54}

{45,46,78}

{98,20,24}

Find max value in complete 2D numpy array :

When we are going to find the max value in a 2D numpy array, we can either do it by finding a single value or we can find column wise or row wise.

CODE:

import numpy as np

arr = np.array([[50, 59, 54], [45, 46, 78], [98, 20, 24]])
# Get the maximum value from the 2D array

maxValue = np.amax(arr)
print("The maximum value inside the array is", maxValue)

Output :
The maximum value inside the array is 98

Column or Row wise value

To find the max value per each row, we can  pass axis =1 and for columns we can use axis =0.

CODE:

import numpy as np

arr = np.array([[50, 59, 54], [45, 46, 78], [98, 20, 24]])
# Get the maximum valuevin rows
maxRows = np.amax(arr, axis=1)
# Get the maximum valuevin columns
maxColumns = np.amax(arr, axis=0)

print(
    "The maximum values in rows are : ",
    maxRows,
    " and the maximum value in columns are : ",
    maxColumns,
)

Output :
The maximum values in rows are :  [59 78 98]  and the maximum value in columns are :  [98 59 78]

Find index of maximum value from 2D numpy array:

CODE:

import numpy as np

arr = np.array([[50, 59, 54], [45, 46, 78], [98, 20, 24]])
# Get the index of max value inside the 2D array
res = np.where(arr == np.amax(arr))
print("Tuple :", res)
print("Now Coordinates of max value in 2D array :")
# zipping both the arrays to find the coordinates
Coordinates = list(zip(res[0], res[1]))

for elem in Coordinates:
    print(elem)


Output :
Tuple : (array([2], dtype=int32), array([0], dtype=int32))
Now Coordinates Of max value in 2D array :
(2, 0)

numpy.amax() & NaN :

amax( ) also propagates the NaN values , which means if there is a NaN value present in the numpy array, then the max value returned by the function will be NaN.

import numpy as np

arr = np.array([[50, 59, np.NaN], [45, 46, 78], [98, 20, 24]])
# amax( ) propagating the NaN values

print("The max element in the numpy array is :", np.amax(arr))

Output :
The max element in the numpy array is : nan