Creating Grid with Numpy Performance

Question

So here is what I try to do,

N=1000
x=np.arange(0,1,1./float(len(N)))    
XX,YY=np.meshgrid(x,x)   

l=len(XX)
grid=np.array([ ([XX[i,i],YY[j,j],0. ]) for i in xrange(l) for j in xrange(l) ])

the numpy routine is rather fast but I need the grid to be in a different form and this takes quite long (I guess because of indexing the numpy array).

Thanks for any suggestions : )

Cheers

len(N) doesn't work by the way

Anake
– Anake

2012-12-11 11:57:51 +00:00
Commented Dec 11, 2012 at 11:57 — Anake
– Anake, Commented Dec 11, 2012 at 11:57

cronos · Accepted Answer · 2012-12-11 08:09:21Z

3

Take advantage of broadcasting:

z = np.zeros([N, N, 3])
z[:,:,0] = x.reshape(-1,1)
z[:,:,1] = x
fast_grid = z.reshape(N*N, 3)

print np.all( grid == fast_grid ) 
True

answered Dec 11, 2012 at 8:09

cronos

2,46819 silver badges18 bronze badges

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

Comments

Bi Rico · Accepted Answer · 2012-12-12 04:05:55Z

0

You could do something like:

grid = np.mgrid[:N, :N, :1]
grid = grid.T.reshape(-1, 3)

answered Dec 12, 2012 at 4:05

Bi Rico

25.9k3 gold badges57 silver badges75 bronze badges

Collectives™ on Stack Overflow

Creating Grid with Numpy Performance

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related