[Scipy-tickets] [SciPy] #1841: RectBivariateSpline fails with nans at the end

SciPy Trac scipy-tickets@scipy....
Fri Feb 8 14:20:00 CST 2013


#1841: RectBivariateSpline fails with nans at the end
-------------------------------+--------------------------------------------
 Reporter:  shie0041           |       Owner:  pv         
     Type:  defect             |      Status:  new        
 Priority:  normal             |   Milestone:  Unscheduled
Component:  scipy.interpolate  |     Version:  0.11.0     
 Keywords:                     |  
-------------------------------+--------------------------------------------
 RectBivariateSpline fails when there are numpy.nan values in the last
 corner of an array.  Interpolation should be possible in an array so long
 as the nan values are sufficiently far away.

 In the attached code, I create a simple 11x11 array called R.  If 2x2
 regions of the four corners are set to nan, interpolation fails even in
 the middle of the array where valid data are present.

 Next I tried to determine if it mattered where the nan values occur.
 Putting nan values in the first three corners still allows the
 interpolation to return a valid result.  nan values in the fourth corner
 cause the result to be nan, even if the other three corners are not nan.

 Matlab's interp2 function with the method set to 'linear' works fine
 regardless of the locations of nans on the corners.

-- 
Ticket URL: <http://projects.scipy.org/scipy/ticket/1841>
SciPy <http://www.scipy.org>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list