CODEWITHSUNDEEP
pythonarraysnumpy
Taylor
Taylor

Reputation: 301

Take average of columns in a numpy array

I have taken data from a csv file using numpy. numpy array has dimensions : 100*20. How do i take average of columns (say col 3,5,8) and replace them with a new column containing average of these 3 cols

If 

   col3 = 1,2,3,4
   col5 = 2,3,4,8
   col8 = 3,4,5,6

then I want to remove these 3 columns and insert a new column in which each entry contains an average of values in these 3 columns

I want to insert a new column: 2,3,4,6, delete previous 3 cols and dimension of final array to be 100*28

Is there any numpy function to do this ?

Upvotes: 5

Views: 20755

Answers (2)

azalea
azalea

Reputation: 12590

a = np.arange(100*30).reshape(100,30) # change this to your own data
cols = [2, 4, 7]                      # columns to calculate averages, i.e. 3,5,8
b = a[:, cols]                        # data of the cols
c = b.mean(axis=1)                    # average
a_no_b = np.delete(a, cols, axis=1)   # data without the cols
a_final = np.c_[a_no_b, c]            # final result

Upvotes: 4

Jakub Bielan
Jakub Bielan

Reputation: 605

For columns axis = 0, so:

X = np.random.randint(0,101,shape=(100,50))
columns_average = X.mean(axis=0)

You get 1D array 'columns_average' with mean for every column, so 50 values.

Upvotes: 2

Related Questions