I need to add a column and a row to an existing Numpy array at a defined position.
-
3This needs a little more informationChuck Vose– Chuck Vose2010-01-04 21:41:10 +00:00Commented Jan 4, 2010 at 21:41
-
2what kind of an array? list of lists, array.array or numpy.array?Antony Hatchkins– Antony Hatchkins2010-01-04 21:42:04 +00:00Commented Jan 4, 2010 at 21:42
Add a comment
|
3 Answers
I assume your column and rows are just a list of lists?
That is, you have the following?
L = [[1,2,3],
[4,5,6]]
To add another row, use the append method of a list.
L.append([7,8,9])
giving
L = [[1,2,3],
[4,5,6],
[7,8,9]]
To add another column, you would have to loop over each row. An easy way to do this is with a list comprehension.
L = [x + [0] for x in L]
giving
L = [[1,2,3,0],
[4,5,6,0]]
3 Comments
Hamish Grubijan
and to convert that to array just do array(lst) ?
Stefan Kendall
That should work. There's probably a better way to do this with numpy, but your original question did not specify such.
ankostis
np.append takes an axis argument specifying the dimension to append; so the way described for adding a new column is discourage because it is not numpy-optimized.There are many ways to do this in numpy, but not all of them let you add the row/column to the target array at any location (e.g., append only allows addition after the last row/column). If you want a single method/function to append either a row or column at any position in a target array, i would go with 'insert':
T = NP.random.randint(0, 10, 20).reshape(5, 4)
c = NP.random.randint(0, 10, 5)
r = NP.random.randint(0, 10, 4)
# add a column to T, at the front:
NP.insert(T, 0, c, axis=1)
# add a column to T, at the end:
NP.insert(T, 4, c, axis=1)
# add a row to T between the first two rows:
NP.insert(T, 2, r, axis=0)
1 Comment
Shreta Ghimire
I am using you solution after a decade and it works perfect for my need. Thanks!
I would suggest sympy Matrix object to do it:
a = [[ 2, 1, 180],
[ 1, 3, 300],
[-1, -4, 0]]
b = [[1,0],
[0,0],
[0,1]]
import sympy as sp
a = sp.Matrix(a).col_insert(-2, sp.Matrix(b))
a.tolist()
Output:
[[2, 1, 0, 1, 180],
[1, 0, 0, 3, 300],
[-1, 0, 1, -4, 0]]
And to continue with a Numpy array, you can use np.asarray(a) instead of a.tolist() (assuming you have imported Numpy as np)
