Passing multiple columns of a 2D array into a lambda function and returning a single column array

Question

I am missing something simple here. I cannot figure out how to pass multiple columns of a 2D array (matrix) and output them as a single column array.

Here is a simple example:

import numpy as np

Z = lambda x1,x2: (x1*x2 - 3)^2 + 1

# a sample 2D array with 15 rows and 2 columns    
x= np.arange(30).reshape((15,2))

answer = [Z(i[0],i[1]) for i in x]

The last line of code is where my problem lies. I would like the output to be a single column array with 15 rows.

As a final note: My code only uses 2 columns as inputs. If it could be further expanded to a flexible number of columns, it would be greatly appreciated.

In the end I don't know what you want to do, but are you aware that you could do the same operation with numpy leaving out the for loop? which would be: (x[:,0]*x[:,1] -3 )^2 + 1. Also x*x would be x**2 instead of x^2 (if you were intentionally doing that please forgive me) — Magellan88
– Magellan88, Commented Apr 1, 2014 at 11:05
Oh and if you really need columns you could do ((x[:,0]*x[:,1] -3 )^2 + 1)[np.newaxis,:] — Magellan88
– Magellan88, Commented Apr 1, 2014 at 11:13

Lee · Accepted Answer · 2014-04-01 09:07:43Z

1

Could you make your last line:

answer = np.array([Z(i[0],i[1]) for i in x]).reshape(15,1)

which gives:

array([[ -2],
       [  0],
       [ 18],
       [ 36],
       [ 70],
       [104],
       [154],
       [204],
       [270],
       [336],
       [418],
       [500],
       [598],
       [696],
       [810]])

answered Apr 1, 2014 at 9:07

Lee

31.4k31 gold badges124 silver badges187 bronze badges

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

2 Comments

Magellan88 Over a year ago

If you wanted to get rid of the 15 (because the size of the array might change) you could also do .reshape(-1,1) or np.array([[Z(i[0],i[1]) for i in x]]) (note the double [[ ]])

Daniel Over a year ago

As a note numpy strives to avoid explicit for loops. Using numpy in this way will actually slow down your calculation.

Magellan88 · Accepted Answer · 2014-04-01 11:16:53Z

1

You could do

import numpy as np

Z = lambda data, i, j: ((data[:,i]*data[:,j] - 3)**2 + 1)[:,np.newaxis]

# a sample 2D array with 15 rows and 2 columns    
x= np.arange(30).reshape((15,2))

answer = Z(x,0,1)

so maybe you don't need the lambda function after all

answered Apr 1, 2014 at 11:16

Magellan88

2,5734 gold badges30 silver badges37 bronze badges

Collectives™ on Stack Overflow

Passing multiple columns of a 2D array into a lambda function and returning a single column array

2 Answers 2

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related