how to add a column into a numpy array

Question

I would like to add a column of '1's on all my rows of first column... For example, before:

9 8 4
3 4 5
3 2 1

after:

1 9 8 4
1 3 4 5
1 3 2 1

My code so far:

import numpy as np
import pandas as pd

training = pd.read_csv('data.csv', header = None)
training
training.shape

Output = (300, 14)

To make data into a matrix:

trainingData = training.values

Insert a column of 1's in the first column on all rows

featureVector = np.insert(featureVector, 0, 1, axis=0)

When I print the variable featureVector, I don't see the 1's...how can I do this or fix it?

Try with axis=1, since you want to change the number of columns. — czr
– czr, Commented Oct 7, 2017 at 21:10

Anton vBR · Accepted Answer · 2017-10-07 21:14:15Z

1

This worked for me:

array = training.values
np.concatenate(([np.ones(len(array))],array.T),axis=0).T

answered Oct 7, 2017 at 21:14

Anton vBR

19k6 gold badges47 silver badges47 bronze badges

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

Comments

RomanPerekhrest · Accepted Answer · 2017-10-07 21:14:53Z

1

With numpy.insert() routine:

import numpy as np

training = np.loadtxt('data.csv', dtype=int)
training = np.insert(training, 0, 1, axis=1)
print(training)

The output:

[[1 9 8 4]
 [1 3 4 5]
 [1 3 2 1]]

answered Oct 7, 2017 at 21:14

RomanPerekhrest

93.1k4 gold badges75 silver badges112 bronze badges

3 Comments

Anton vBR Over a year ago

So basically he/she had the axis wrong. Nice one! +1

KatZee Over a year ago

just curious, why do you use np.loadtxt() over pd.read_csv? can np.loadtxt() handle large data sets?

RomanPerekhrest Over a year ago

@KatZee, just seen that simple input data and thought that the job could be done with single numpy module(without pandas). Of course, you may continue to use pandas

Eric Duminil · Accepted Answer · 2017-10-07 21:20:40Z

0

You could concatenate a 1-column matrix and training :

>>> import numpy as np
>>> training = np.array([9, 8, 4, 3, 4, 5, 3, 2, 1]).reshape(3, 3)
>>> np.concatenate((np.ones((training.shape[1], training)), a), axis=1)
array([[ 1.,  9.,  8.,  4.],
       [ 1.,  3.,  4.,  5.],
       [ 1.,  3.,  2.,  1.]])

axis=1 is used to specify that a column should be added. Otherwise:

>>> np.concatenate(([[1,1,1]], training), axis=0)
array([[1, 1, 1],
       [9, 8, 4],
       [3, 4, 5],
       [3, 2, 1]])

You could also use column_stack:

>>> np.column_stack((np.ones(training.shape[1]), training))
array([[ 1.,  9.,  8.,  4.],
       [ 1.,  3.,  4.,  5.],
       [ 1.,  3.,  2.,  1.]])

edited Oct 7, 2017 at 21:20

answered Oct 7, 2017 at 21:13

Eric Duminil

54.6k10 gold badges80 silver badges134 bronze badges

1 Comment

KatZee Over a year ago

would I have to make a loop if I wanted to change 300 rows in the first column?

Collectives™ on Stack Overflow

how to add a column into a numpy array

3 Answers 3

Comments

3 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

3 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related