2

I would like to use an equivalent to scipy griddata because of an error appears : 

QH7074 qhull warning: more than 16777215 ridges.  ID field overflows and two ridges
may have the same identifier.  Otherwise output ok.

which sometimes kill my calculation or just very slow...

Currently i have :

test_interp= griddata((xx_points.flatten(),yy_points.flatten()),values.flatten(), (xi+X,yi+V), method='linear',fill_value=nan)
user3601754
  • 3,792
  • 11
  • 43
  • 77
  • I think it s due to the importance of value to treat... :/ – user3601754 Mar 06 '15 at 10:10
  • 1
    If your original data is on a rectangular grid, see https://stackoverflow.com/questions/26276550/alternatives-to-python-griddata/26282370 – pv. Mar 06 '15 at 11:34

1 Answers1

0

I had exactly the same error message and it looks like a limitation of griddata for big data, according to this thread. The limit is 2^24=16777216 'ridges' (I don't know exactly what it refers to). I did exactly the same interpolation with less data points and this problem didn't occur. My guess is that you should find a way to split your dataset to be below the 2^24 threshold.

Bouliech
  • 84
  • 1
  • 3