sparse 2-D lookup

[Jeff Abrahamson]

I want an array of (x,y,z) triples.  The problem is that x and y come
from the same very large set, but very few (x,y) pairs will actually
be used.  The z value is computed based on x and y, so this is really
a function f(x,y) = z.

Naively I'd create a sparse array, but sparse arrays seem (it's not
clear, though) to become regular arrays if I add elements to them.
Thus,

    v = sparse(10000000,10000000)
    v =

    Compressed Column Sparse (rows=10000000, cols=10000000, nnz=0)

    octave:3> v(1,1) = 1 

(and memory usage jumps to near 2 GB and my machine slows to a crawl).

Any suggestions on how to do the bookkeeping?


The particular task is that I want to subdivide a triangulated surface
using midpoints of edges.  So I have a 3 x n array vertices and a 3 x n
of triangles.  I want to add midpoints to the vertex array, but I
don't want to add each midpoint twice, so I want an efficient lookup
mechanism so I can say "Is the midpoint of (x,y) already assigned?  If
yes, what is its index in the vertex array, if no add it for me."

Thanks much for any pointers.

-- 
 Jeff

 Jeff Abrahamson  <http://www.purple.com/jeff/>    +1 215/837-2287
 GPG fingerprint: 1A1A BA95 D082 A558 A276  63C6 16BF 8C4C 0D1D AE4B



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------