[Numpy-discussion] Faster

[Robert Kern]

On Fri, May 2, 2008 at 7:24 PM, Keith Goodman <kwgoodman@gmail.com> wrote:
> How can I make this function faster? It removes the i-th row and
>  column from an array.
>
>  def cut(x, i):
>     idx = range(x.shape[0])
>     idx.remove(i)
>     y = x[idx,:]
>     y = y[:,idx]
>     return y
>
>  >> import numpy as np
>  >> x = np.random.rand(500,500)
>  >> timeit cut(x, 100)
>  100 loops, best of 3: 16.8 ms per loop

I can get a ~20% improvement with the following:

In [8]: %timeit cut(x, 100)
10 loops, best of 3: 21.6 ms per loop

In [9]: def mycut(x, i):
   ...:     A = x[:i,:i]
   ...:     B = x[:i,i+1:]
   ...:     C = x[i+1:,:i]
   ...:     D = x[i+1:,i+1:]
   ...:     return hstack([vstack([A,C]),vstack([B,D])])
   ...:

In [10]: %timeit mycut(x, 100)
10 loops, best of 3: 17.3 ms per loop

The hstack(vstack, vstack) seems to be somewhat better than
vstack(hstack, hstack), at least for these sizes.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless
enigma that is made terrible by our own mad attempt to interpret it as
though it had an underlying truth."
 -- Umberto Eco