Transform from 3dimension to 4dimension

[python/iris]

HI all,

I have 3dimesion precipitation cubes

like this,

>>> cubes = iris.load('*.nc') 

>>> print cubes

cubes[0] =     time : 360, latitude: 145, longitude:192

cubes[1] =     time : 360, latitude: 145, longitude:192

cubes[2]) =     time : 360, latitude: 145, longitude:192

cubes[3] =     time : 360, latitude: 145, longitude:192

cubes[4]=     time : 360, latitude: 145, longitude:192

5 cubes no scalar coordinate. therefore, merge function is not working. 

i attempted to slove the problem, like this

>>> ncube = iris.util.new_axis(cubes[0], 'time')

>>> print cubes[0]

cubes[0] =                        --:1  time:360 , latitude:145, longitude:192

dimension coordinates:      

time                                    -        x                -                      -

latitude                               -        -                 x                     -

longitude                            -        -                 -                     x

  conventions: CF-1.5

>>> ncube.add_dim_coord(iris.coords.DimCoords([1], 'height'),0)

>>> print ncube

cubes[0] =                        height:1  time:360 , latitude:145, longitude:192

dimension coordinates:      

height                                 x        -                -                      -

time                                    -        x                -                      -

latitude                               -        -                 x                     -

longitude                            -        -                 -                     x

is this right?

Ultimately, I want to multi mean ensemble..

Please advise me what to do

[Richard Hattersley]

As you say, the merge and merge_cube methods work with scalar coordinates. But a scalar coordinate isn't associated with a dimension. It looks like what you want to do is something like:

for i, cube in enumerate(cubes):
    cube.add_aux_coord(iris.coords.DimCoord(i, 'height'))

merged_cube = cubes.merge_cube()

Did you really mean to use "height" to label all the cubes? Should that have been "realization" instead?

[python/iris]

Hi Richard.

Thank you for your answers.

yes, I want to cubes has 'realization' coordinate..

and i tried,

fileform =  pr_day_MIROC5_rcp85_r1i1p1_2007_2049_ensemble_001.nc

----------------------------------------------------------------------------------

import iris

import numpy as np

def realization_metadata(cube, field, fname):

       if not cube.coords('realization'):
              realization_number = fname[-6:-3]

              import iris.coords

              realization_coord = iris.coords.AuxCoord(np.int32(realization_number), 'realization')

              cube.add_aux_coord(realization_coord)

filenmae = iris.sample_data_path('/2007-2049/*')

cubes  = iris.load(filename(callback=realization_metadata)

cubes = iris.cube.CubeList([cube[:,:,:] for cube in cubes])

newcube = cubes.merge_cube()

print newcube

-------------------------------------------------------------------------------------------

but two error

newcube = cubes.merge_cube()

iris.exception.MergeError: failed to merge into a single cube.

cube.shape differs: (12.360.145.192) !=(360,,145,192)               # i think, it is not problems.

cube data fill_value differs:9.969209968386869e+36 !=1e+20

and

cube data dtype differs: float 64 != float 32

i think, I use the "iris.fileformat.netcdf.NetCDFDataProxy"   but i'm not sure.

how do i slove this problem..