Error when putdata with an pyferret.AXISTYPE_ABSTRACT axis

[PBrockmann]

I have encountered a problem when you want to pass a python object to ferret and want to work with an abstract axis to stay simple.

Here, I have simplified the problem by getting and putting back the same object. It fails when the object has an abstract axis. I also have found a solution by writing in the axis_coords tuple a None rather than the values themselves.

Otherwise you get this error:

`

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

`

Here is the code to test this:

import numpy as np
import pyferret

#========================================================
pyferret.start()
pyferret.run('use monthly_navy_winds.cdf')

#========================================================
# Simple get and put same data ---> works !

a = pyferret.getdata('uwnd', False)

a['name']='myvar'
print a.keys()
print a['axis_types']

pyferret.putdata(a,None)
pyferret.run('show data')

#========================================================
# Change the variable to an abstract axis 
# then get and put it back for testing --> fails !

pyferret.run('def axis/t=1:132:1 mytaxis')
pyferret.run('let uwnd2 = uwnd[gt=mytaxis@ASN]')
pyferret.run('show grid uwnd2')

a = pyferret.getdata('uwnd2', False)
a['name']='myvar2'
print a.keys()

a['axis_coords'] = (
                    a['axis_coords'][0],
                    a['axis_coords'][1],
                    a['axis_coords'][2],
                    a['axis_coords'][3],      # <--- to None to get rid off the error
                    a['axis_coords'][4],
                    a['axis_coords'][5],
)

pyferret.putdata(a,None)
pyferret.run('show data')

Error:

/opt/pyferret-7.4.4-RHEL7-64-Python2.7/lib/python2.7/site-packages/pyferret/datamethods.pyc in putdata(datavar_dict, axis_pos)
    541                 raise ValueError("number of coordinates for axis %d does not match the number of data points" % (k+1))
    542         elif axis_types[k] == libpyferret.AXISTYPE_ABSTRACT:
--> 543             if axis_coords[k] != None:
    544                 axis_coords[k] = numpy.array(axis_coords[k], dtype=numpy.float64, copy=1)
    545                 if axis_coords[k].shape[0] != shape[k]:

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()