Find all maximums on a numpy array

Question

I want to find all the maximums and their index in a numpy array.

I am looking for something simple. I started something like data[1:] < data[:-1] no idea how to get 20 and 30 and their index 2 and 7.

import numpy as np

data = np.array([0,1,20,1,0,1,2,30,2,1,0])

A = data[1:] < data[:-1]

EDIT: Argmax and findpeaks sound difficult (find peaks gives you min and max for instance). I did a np.diff and a loop and I have my index eventhough I try to avoid using loop

import numpy as np

def maximum_index(data):  
    data = np.diff(data)
    index_max = []
    for index in range(len(data)-1):
        if data[index] > 0 and data[index+1] < 0:
            index_max.append(index+1)
    return np.asarray(index_max)


data = np.array([0,1,9,1,0,1,2,9,2,1,0])
result_max = maximum_index(data)

Thank you

Does this answer your question? [How does numpy argmax work?](https://stackoverflow.com/questions/50460994/how-does-numpy-argmax-work) — Joe, Apr 22 '20 at 08:08
https://stackoverflow.com/questions/40357335/numpy-how-to-get-a-max-from-an-argmax-result — Joe, Apr 22 '20 at 08:08

score 2 · Answer 1 · edited Apr 22 '20 at 09:04

You can use numpy.argsort as follows:

import numpy as np

data = np.array([0,1,20,1,0,1,2,30,2,1,0])
N = 2 # number of maxima

idx = np.argsort(data)[-N:]
idx
array([2, 7], dtype=int64)
data[idx]
array([20, 30])

If you don't know the number of maxima, you can use scipy.signal.find_peaks:

from scipy.signal import find_peaks

idx = find_peaks(data)[0]
idx
array([2, 7], dtype=int64)
data[idx]
array([20, 30])

score 0 · Answer 2 · answered Apr 22 '20 at 07:36

you can try below code

max_data_pos=np.argmax(data)        #Find position of max elelemt
max_data_val=data[max_data_pos]     #Find Max Valuse 

second_max_data_pos=np.argmax(np.delete(data,max_data_pos))  #Find position of second max elelemt
secondmax_data_val=data[second_max_data_pos]                 #Find Max Valuse

Find all maximums on a numpy array

2 Answers2