[GRASS-dev] [GRASS GIS] #2466: r.param.scale slow on Windows when there is no data in raster

[GRASS GIS]

#2466: r.param.scale slow on Windows when there is no data in raster
--------------------------------+-------------------------------------------
 Reporter:  annakrat            |       Owner:  grass-dev@…              
     Type:  defect              |      Status:  new                      
 Priority:  normal              |   Milestone:  7.0.0                    
Component:  Raster              |     Version:  unspecified              
 Keywords:  r.param.scale, nan  |    Platform:  MSWindows 8              
      Cpu:  Unspecified         |  
--------------------------------+-------------------------------------------

Comment(by glynn):

 Replying to [comment:2 annakrat]:

 > It's not clear to me, what the algorithm should do when there is a null
 value at one of the cells in the search window. It could just skip the
 entire computation and write null in the central cell. This would be
 probably the most correct option but it results in null edges. Or we could
 assign 0 values but then the result on edges would be distorted by the
 zeros. Which approach should we implement?

 AFAICT, you either need to accept that nulls will propagate (i.e. if any
 cell in the window is null, the result will be null), or replace the
 least-squares fitting algorithm with something which can handle nulls.

 FWIW, that module dates back to at least GRASS 4.x, when neither nulls nor
 floating-point existed.

-- 
Ticket URL: <http://trac.osgeo.org/grass/ticket/2466#comment:3>
GRASS GIS <http://grass.osgeo.org>