How do I combine rowwise values of 2D Numpy array into one single column

Question

I have a 2D numpy array as shown below

[['me' 'you']
 ['her' 'him']]

I want to get a numpy array like this

['meyou'
'herhim']

which should be a 1D numpy array

How do I do it in the most efficient way in terms of speed?

How ever you do it, it's a python string operation. Even with np.char. So there's no superfast array approach. — hpaulj
– hpaulj, Commented Aug 11, 2018 at 0:13

John Zwinck · Accepted Answer · 2018-08-10 23:48:58Z

1

Simply "add" the two columns using NumPy's string manipulation:

 np.char.add(x[:,0], x[:,1])

Or more generally, for any number of columns, here is a simple solution (if the number of columns is large, this is not efficient):

from functools import reduce # not needed in Python 2
reduce(np.char.add, x.T)

answered Aug 10, 2018 at 23:48

John Zwinck

252k44 gold badges346 silver badges459 bronze badges

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

Comments

Onyambu · Accepted Answer · 2018-08-11 00:18:32Z

1

You can use a list comprehension, whereby you join all the values within the same sublist together:

["".join(i) for i in x]

['meyou', 'herhim']

answered Aug 11, 2018 at 0:18

Onyambu

80.3k3 gold badges29 silver badges65 bronze badges

Even if you take into account casting this back into a numpy array it's faster than using numpy methods