How to create a netCDF file with python netCDF4?

Question

I am learning how to use netCDF4 using Python module in Anaconda. I am trying to append values to two variables I have created time and field:

from netCDF4 import Dataset
import numpy as np

root_grp = Dataset('py_netcdf4.nc', 'w', format='NETCDF4')
root_grp.description = 'Example simulation data'

ndim = 128 # Size of the matrix ndim*ndim
xdimension = 0.75
ydimension = 0.75
# dimensions
root_grp.createDimension('time', None)
root_grp.createDimension('x', ndim)
root_grp.createDimension('y', ndim)

# variables
time = root_grp.createVariable('time', 'f8', ('time',))
x = root_grp.createVariable('x', 'f4', ('x',))
y = root_grp.createVariable('y', 'f4', ('y',))
field = root_grp.createVariable('field', 'f8', ('time', 'x', 'y',))

# data
x_range =  np.linspace(0, xdimension, ndim)
y_range =  np.linspace(0, ydimension, ndim)
x[:] = x_range
y[:] = y_range
for i in range(5):
    time[i] = i*50.0
    field[i,:,:] = np.random.uniform(size=(len(x_range), len(y_range)))


root_grp.close

But now when I print one of the variables I get that it's empty(!!):

Python 2.7.10 |Anaconda 2.4.1 (64-bit)| (default, Sep 15 2015, 14:50:01) 
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
Anaconda is brought to you by Continuum Analytics.
Please check out: http://continuum.io/thanks and https://anaconda.org
>>> from netCDF4 import Dataset
>>> root_grp = Dataset('py_netcdf4.nc', 'r', format='NETCDF4')
>>> print root_grp.variables["field"][:,:,:]
[]
>>> 

What am I doing wrong here?

Solution

This works:

from netCDF4 import Dataset
import numpy as np

root_grp = Dataset('py_netcdf4.nc', 'w', format='NETCDF4')
root_grp.description = 'Example simulation data'

ndim = 128 # Size of the matrix ndim*ndim
xdimension = 0.75
ydimension = 0.75
# dimensions
root_grp.createDimension('time', None)
root_grp.createDimension('x', ndim)
root_grp.createDimension('y', ndim)

# variables
time = root_grp.createVariable('time', 'f8', ('time',))
x = root_grp.createVariable('x', 'f4', ('x',))
y = root_grp.createVariable('y', 'f4', ('y',))
field = root_grp.createVariable('field', 'f8', ('time', 'x', 'y',))

# data
x_range =  np.linspace(0, xdimension, ndim)
y_range =  np.linspace(0, ydimension, ndim)
x[:] = x_range
y[:] = y_range
for i in range(5):
    time[i] = i*50.0
    field[i,:,:] = np.random.uniform(size=(len(x_range), len(y_range)))


root_grp.close()

The only difference is that I call the close() method: root_grp.close().

Comments

[j08lue]

To avoid mistakes like that (e.g. missing brackets on close method), use the python with statement, like in this example.