# pasting numpy array into bigger array

Discussion in 'Python' started by TG, Jul 26, 2006.

1. ### TGGuest

hi.

let's say I have :

from numpy import *
x = identity(5)
y = zeros((7,7))

I want to paste x into y, starting at coordinates (1,1) in order to
change y to something like this :

0 0 0 0 0 0 0
0 1 0 0 0 0 0
0 0 1 0 0 0 0
0 0 0 1 0 0 0
0 0 0 0 1 0 0
0 0 0 0 0 1 0
0 0 0 0 0 0 0

how would you do that ?

TG, Jul 26, 2006

2. ### Tim HeaneyGuest

"TG" <> writes:
>
> let's say I have :
>
> from numpy import *
> x = identity(5)
> y = zeros((7,7))
>
> I want to paste x into y, starting at coordinates (1,1) in order to
> change y to something like this :
>
> 0 0 0 0 0 0 0
> 0 1 0 0 0 0 0
> 0 0 1 0 0 0 0
> 0 0 0 1 0 0 0
> 0 0 0 0 1 0 0
> 0 0 0 0 0 1 0
> 0 0 0 0 0 0 0
>
> how would you do that ?

You can use Python slice notation for each dimension

y[1:6,1:6] = x

In general, I guess you want something like

y[1:1+x.shape[0], 1:1+x.shape[1]] = x

or

m, n = 1, 1
s, t = x.shape
y[m:m+s, n:n+t] = x

There is a mailing list for numpy

https://lists.sourceforge.net/lists/listinfo/numpy-discussion

I think it's a shame there isn't any free documentation for numpy.

Tim

Tim Heaney, Jul 26, 2006

3. ### TGGuest

Thanks, that's exactly what I needed.

Tim Heaney wrote:
> You can use Python slice notation for each dimension
>
> y[1:6,1:6] = x
>

TG, Jul 26, 2006
4. ### Robert KernGuest

Robert Kern, Jul 26, 2006