Numpy aggregate functions max to get highest of elements with different options

numpy.max(a,axis=None,out=None,keepdims, initial, where)

Return highest of elements across given axis.

`a`	array, elements to get the max value
`axis`	Int (optional ), or tuple, default is None, maximum value among all the elements. If axis given then values across the axis is returned.
`out`	Optional. If given then output to be stored. Must be of same shape as of the output
`keepdims`	Bool ( Optional ), output matches to the input array dimension.
`where`	Optional, Elements to include for calculation of maximum value
`initial`	Optional, int, Minimum value of the output. If given, then this is considered if it is more than the actual output

We will use these parameters in our examples.

Sample array

You can use randint() to create an array for our examples. Or can use fixed elements to create the array.

import numpy as np
# my_data=np.random.randint(2,high=7,size=(3,3),dtype='int16') 
my_data=np.array([[6, 3, 2], [7, 2, 2], [6, 2, 9]])
print(my_data)

Output

[[6 3 2]
 [7 2 2]
 [6 2 9]]

Axis

Maximum vlaue of the elements across the axis. Axis of Numpy max array

print("max()      : ", my_data.max())
print("max(axis=0): ", my_data.max(axis=0))
print("max(axis=1): ", my_data.max(axis=1))

Output

max()       :  9
max(axis=0) :  [7 3 9]
max(axis=1) :  [6 7 9]

out

Alternative output array, must be of same shape as expected output. Let us first check with axis.

x = np.zeros(3,dtype=int)
print(my_data.max(axis=0,out=x))
print(x)

Output

[7 3 9]
[7 3 9]

Without using axis

y = np.array(1)
print(my_data.max(out=y))
print(y)

Output

9
9

keepdims

If it is set to True ( keepdims=True ) then it will take the dimension of input array.

print("max(keepdims=True) : ", my_data.max(keepdims=True))
print("max(keepdims=False) : ", my_data.max(keepdims=False))

Output

max(keepdims=True) :  [[9]]
max(keepdims=False) :  9

Using where

By using where we can say which elements to use and which elements not to use ( by setting True or False ) . While using where we have to give initial value.

print(my_data.max(where=[True, False,True],initial=2))

Output

Using axis with where

print(my_data.max(axis=1,where=[True, False,True],initial=2))
print(my_data.max(axis=0,where=[True, False,True],initial=2))

Output

[6 2 6]
[6 2 3]

initial

We can assign initial value to our output. Note that the final output will be maximum of initial vlaue and actual min value ( without the initial value ).

print(my_data.max())  # 9
print(my_data.max(initial=2)) # 9
print(my_data.max(initial=12)) # 12

Output

9
9
12

Check the where option above. The value assigned to intial value is given as output where it is less than the actual output.

Intial vlaue with Axis option

Compare the values with initial value and without intitial value. We used axis=0 , you can try with axis=1.

print(my_data.max(axis=0))  # [7 3 9]
print(my_data.max(axis=0,initial=2)) # [7 3 9]
print(my_data.max(axis=0,initial=12)) # [12 12 12]

Output

[7 3 9]
[7 3 9]
[12 12 12]

Numpy mean() sum() min()

Python Tkinter Pandas Python- Tutorials

Subhendu Mohapatra

Author & Instructor at plus2net

I write and maintain practical tutorials on Python, PHP, SQL, JavaScript, HTML, jQuery, and web development at plus2net. The tutorials focus on clear explanations, working examples, and code that readers can test and adapt while learning.

About the author Watch on YouTube

Connect:

Subscribe to our YouTube Channel here

Numpy max()

Sample array

Axis

out

keepdims

Using where

initial

Intial vlaue with Axis option

Subhendu Mohapatra