Re: FIXED (again): INTERP2 problem

[Paul Kienzle]

On Nov 26, 2005, at 1:43 PM, Pierre Baldensperger wrote:

> As mentioned in the previous posts: the behaviour of this version of 
> INTERP2
> with respect to XI and YI has been changed!! So if your code was 
> assuming the
> former behaviour (imposed meshgridded XI and YI), you may need to 
> modify the
> way you call the function.

I'm happy to support XI and YI arbitrary, however according to the docs
on the web, interp2(z,[1:5],[1:6]) should return a 5x6 matrix. Do you 
have
reason to believe otherwise?  I've reverted the behaviour to 
automatically
generate a meshgrid if given vectors.

Please add a test for non-gridded XI,YI at the bottom of the file.

I've updated your patch to include recent changes on CVS (in particular
bicubic interpolation) and to add a test case for interp2(z,n).  bicubic
doesn't handle non-meshgridded data, and silently does the old behaviour
of taking the first column of XI and row of YI.

Please perform future revisions against the new version.

Thanks,

        - Paul

interp2.m
Description: Binary data