CDAT does not handle unicode string numpy datatypes

[aashish24]

Which I think is odd :( we should have support for both of them.

[doutriaux1]

What are you trying to do? Can you post an example?

[aashish24]

Very simple passing a unicode string to the cdms2 open call and variable call and then passing a coordinate to retrieve the value something like this:

f = cdms2.open(unicode_filename_str)
v = f[unicode_varname_str]
....

[doutriaux1]

ok so you meant cdms2 not cdat. Easy enough to fix I'll push a branch today.

[aashish24]

is cdms2 is not part of cdat -:) just wanted to make sure that I call them correctly. In general people separate these two things.

Thanks that would be great. Now a days most of the web work is using unicode string (for obvious reasons) and numpy data types are used because we get them from cdat.

[doutriaux1]

it just helps to pinpoint the issue if we know what doesn't work ;)

[doutriaux1]

The following works:

import cdms2
import sys
fnm= u"%s/sample_data/clt.nc"%sys.prefix
print fnm
f=cdms2.open(fnm)
var=u"clt"
print var
s=f[var]
print s.shape

Can you copy/paste an example that doesnt work

[aashish24]

Sure, we have that example from the web work. I will see if I can copy paste here.

d = self._canvas.plot(data,self._plotTemplate,self._plotType,bg=1)

In this line, self._plotType had to be a non-unicode str.

[doutriaux1]

ok so you can confirm cdms2 works, but it is in vcs? I should be able to push a fix today.
By the way, you probably shouldn't use _plotType, var starting with an underscore are usually meant to be private.

[doutriaux1]

seems to be duplicte #537