[Scipy-tickets] [SciPy] #898: bug with interp2d

SciPy Trac scipy-tickets@scipy....
Thu Nov 22 16:51:41 CST 2012


#898: bug with interp2d
--------------------------------+-------------------------------------------
  Reporter:  jlm                |       Owner:  pv              
      Type:  defect             |      Status:  closed          
  Priority:  normal             |   Milestone:  Unscheduled     
 Component:  scipy.interpolate  |     Version:  0.7.0           
Resolution:  wontfix            |    Keywords:  interp2d, linear
--------------------------------+-------------------------------------------
Changes (by pv):

  * status:  new => closed
  * resolution:  => wontfix


Comment:

 I looked at this again: the issue appears to be that the knot selection in
 FITPACK is not necessarily symmetric around the center even if the data
 is. The 'linear' interpolation in FITPACK in any case means that the
 interpolation is linear between knots (which are selected to lie somewhere
 in the region), and not between data points.

 I don't think we want to modify FITPACK to enforce such symmetry. So, I
 don't think there is anything to fix in the present behavior, although at
 first sight it does seem rather strange.

 However, in the future interp2d probably will use a more robust approach
 for interpolation when the original data is gridded, which does not have
 these problems: https://github.com/scipy/scipy/pull/353

-- 
Ticket URL: <http://projects.scipy.org/scipy/ticket/898#comment:4>
SciPy <http://www.scipy.org>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list