cdms2.getGrid()/setGrid() doesn't work

[durack1]

(Pdb) print s.getGrid()
<CurveGrid, id: grid_37, shape: (384, 320)>
(Pdb) heatContent.setGrid(s.getGrid())
*** CDMSError: grid does not match axes for variable variable_524300
(Pdb) heatContent.shape
(60, 384, 320)
(Pdb) print s.shape
(60, 384, 320)
(Pdb) print s.getGrid().shape
(384, 320)
(Pdb) print cdm.createGenericGrid(s.getLatitude(),s.getLongitude(),order='yx')
*** ValueError: could not broadcast input array from shape (384,320) into shape (384)
(Pdb) print cdm.createGenericGrid(s.getLatitude(),s.getLongitude(),order='xy')
*** ValueError: could not broadcast input array from shape (384,320) into shape (384)
(Pdb) print s.getGrid().clone()
<CurveGrid, id: grid_37, shape: (384, 320)>
(Pdb) heatContent.setGrid(s.getGrid().clone())
*** CDMSError: grid does not match axes for variable variable_524300

[chaosphere2112]

@dnadeau4 @durack1 createGenericGrid is a lie; it's actually creating a rectangular grid. The reason setGrid doesn't work is because clone() instantiates new axes, and the setGrid() function checks that the axes on the grid are the same python objects as the ones on the variable. I feel like just making sure they're the same shape should be enough, but no scientist am I.

[durack1]

@chaosphere2112 I'm not sure I get all the subtleties in the above, however in addition to a shape check a min/max bounds check of both grid lats/lons would be a good way to sanity check things.. happy to provide more in depth feedback if you like

[dnadeau4]

@durack1 can you give me a file with these grids.

[dnadeau4]

46b27f5cc9314da268119dee2b2d15115463439e

[dnadeau4]

import cdms2
import numpy
f=cdms2.open("/software/anaconda3/envs/cdms3/share/uvcdat/sample_data/so_Omon_ACCESS1-0_historical_r1i1p1_185001-185412_2timesteps.nc")
f.listvariables()
data=f('so')

grid=data.getGrid()
bb=numpy.zeros((300,360))
bb=cdms2.tvariable.TransientVariable(bb)
bb.setGrid(cdms2.hgrid.TransientCurveGrid(data.getLatitude(), data.getLongitude()))
curvgrid=cdms2.hgrid.TransientCurveGrid(data.getLatitude(), data.getLongitude())

[durack1]

@dnadeau4 great, so all fixed? I assume you've run this across a couple of the test data - the so_Omon* file should have a bunch of others alongside it.