[Numpy-discussion] Finding a row match within a numpy array

Stefan van der Walt stefan@sun.ac...
Sat Aug 18 05:25:21 CDT 2007


On Tue, Aug 14, 2007 at 11:53:03AM +0100, Andy Cheesman wrote:
> Dear nice people
> 
> I'm trying to match a row (b) within a large numpy array (a). My most
> successful attempt is below
> 
> hit = equal(b, a)
> total_hits = add.reduce(hit, 1)
> max_hit = argmax(total_hits, 0)
> answer = a[max_hit]
> 
> where ...
> a = array([[ 0,  1,  2,  3],
>     	   [ 4,  5,  6,  7],
> 	   [ 8,  9, 10, 11],
> 	   [12, 13, 14, 15]])
> 
> b = array([8,  9, 10, 11])
> 
> 
> 
> I was wondering if people could suggest a possible more efficient route
> as there seems to be numerous steps.

For large arrays, you may not want to calculate a == b, so you could
also do

[row for row in a if N.all(row == b)]

or find the indices using

[r for r,row in enumerate(a) if N.all(row == b)]

Cheers
Stéfan


More information about the Numpy-discussion mailing list