How to append a set of numpy arrays while inside a for loop?

Question

The thing I want to do is, I want append a set of numpy arrays one after the other while in a for loop. So whenever a new array comes, the new array will be appended to the older one. The code I have written is as follows:

def joinArray(highest_score, seq):

      seq1 = np.array([])
      seq2 = np.array([])
      seq3 = np.array([])
      seq4 = np.array([])
      seq5 = np.array([])

      if(highest_score == 1):
    
            seq1 = np.concatenate((seq1,seq), axis = 1)
    
            return(seq1)

      if(highest_score == 2):
    
            seq2 = np.concatenate((seq2,seq), axis = 1)

      if(highest_score == 3):
    
            seq3 = np.concatenate((seq3,seq), axis = 1)

      if(highest_score == 4):
    
            seq4 = np.concatenate((seq4,seq), axis = 1)

      if(highest_score == 5):
    
            seq5 = np.concatenate((seq5,seq), axis = 1)

def sequenced(dat):

      for i in range(0,10):
    
          passvalue = joinArray(highest_index[i]+1,dat)



for i in range(0,38,2):

       H = Data2.iloc[:, i:i+2]
       H = H.dropna()

       scaler = StandardScaler()
       Sdatad1 = scaler.fit_transform(H)

h1 = sequenced(Sdatad1)

However, on running this code, I get this error:

   AxisError: axis 1 is out of bounds for array of dimension 1

I have tried np.append(), np.vstack() also but all end up in errors. Is there a way this problem can be solved?

np.concatenate requires careful attention to argument shapes. What the shape of seq1, or seq? axis=1 requires at least 2 dimensions. You can append elements to a list without much thought as to the contents, since it just adds a reference in-place. concatenate (and derivatives) makes a whole new array, and must align shapes just right. It's making a multidimensional array, not just taking on a reference. — hpaulj
– hpaulj, Commented Jul 25, 2021 at 16:19

chazecka · Accepted Answer · 2021-07-25 13:41:37Z

1

First of all, you are redefining your arrays each run. I used a little trick with global keyword here, but you can easily convert the code into OOP, if you are familiar with it. Secondly, in order for python to append array to array (and not number after number), you have to define your array as 1- dimension array. I simplified your joinArray function, but you should be able to continue working with it.

import numpy as np

highest_score_HMM = 1

seq1 = np.empty((0, 1))
seq2 = np.empty((0, 1))
seq3 = np.empty((0, 1))
seq4 = np.empty((0, 1))
seq5 = np.empty((0, 1))


def joinArray(highest_score, seq):

    global seq1, seq2, seq3, seq4, seq5

    if(highest_score == 1):
        seq1 = np.concatenate((seq1, seq))

    if(highest_score_HMM == 2):
        seq2 = np.concatenate(seq2, seq)

    if(highest_score_HMM == 3):
        seq3 = np.concatenate(seq3, seq)

    if(highest_score_HMM == 4):
        seq4 = np.concatenate(seq4, seq)

    if(highest_score_HMM == 5):
        seq5 = np.concatenate(seq5, seq)


seq = np.array([[2]])
joinArray(1, seq)
joinArray(1, seq)
joinArray(1, seq)

print(seq1)

result: [[2.][2.][2.]]

result[0]: [2.]

answered Jul 25, 2021 at 13:41

chazecka

1519 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

sumitpal0593 Over a year ago

Thanks! Its working but unfortunately as my arrays have different dimensions, appending them gives an error.

hpaulj Over a year ago

Repeated concatenate is slower than list append, and requires careful attention to the array shape, both as initialed and in the loop. Each concatenate makes a whole new array, with all the implied copying.

Collectives™ on Stack Overflow

How to append a set of numpy arrays while inside a for loop?

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related