Unmasking of masked Numpy array changes masked values to 0's

Question

I mask my array where values are nodata (-9999), calculate the mean on axis = 0 and then unmask my data array, but then my nodata values are changed into 0's, but now how to make a distinction between "calculated mean 0's" and "nodata 0's". See following code example:

In [1]: import numpy.ma as ma
   ...: x = [[0.,1.,-9999.,3.,4.],[0.,2.,-9999,4.,5.]]
   ...: x 
Out[1]: [[0.0, 1.0, -9999.0, 3.0, 4.0], [0.0, 2.0, -9999, 4.0, 5.0]]

In [2]: mx = ma.masked_values(x, -9999.)
   ...: mx
Out[2]: 
masked_array(data =
 [[0.0 1.0 -- 3.0 4.0]
 [0.0 2.0 -- 4.0 5.0]],
             mask =
 [[False False  True False False]
 [False False  True False False]],
       fill_value = -9999.0)

In [3]: mean = mx.mean(axis=0)
   ...: mean
Out[3]: 
masked_array(data = [0.0 1.5 -- 3.5 4.5],
             mask = [False False  True False False],
       fill_value = 1e+20)

In [4]: mean.mask = ma.nomask
   ...: mean
Out[4]: 
masked_array(data = [0.0 1.5 0.0 3.5 4.5],
             mask = [False False False False False],
       fill_value = 1e+20)

But I would like to have an output similar to my input, with nodata values as -9999., like:

In [4]: mean.mask = ma.nomask
   ...: mean
Out[4]: 
masked_array(data = [0.0 1.5 -9999. 3.5 4.5],
             mask = [False False False False False],
       fill_value = 1e+20)

Answer 1

>>> mean = mx.mean(axis=0)
>>> mean[mean.mask] = mx.fill_value
>>> mean
masked_array(data = [0.0 1.5 -9999.0 3.5 4.5],
             mask = [False False False False False],
       fill_value = 1e+20)