Ticket #31 (defect)

In each proc_*.py creator module, change dimension names to match corresponding variable name. It might not make reading the file by humans easy, but it will allow pydap to work correctly to grab data dependency.

For example, pydap cannot load d2.wtemp[:] because ntime dimension does not match time var. pydap is looking for the dependency based on dimension values.

>>> d2= pydap.open_url('http://whewell.marine.unc.edu/dods/nccoos/level1/b2/ctd1/b2_ctd1_2011_11.nc')             
>>> d2.wtemp
{'wtemp': <pydap.model.BaseType object at 0xb792fb8c>, 'ntime': <pydap.model.BaseType object at 0xb792fc2c>}

>>> d2.wtemp[:]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/opt/env/haines/dataproc/lib/python2.6/site-packages/pydap/model.py", line 662, in __getitem__
    var.data = var.data[slice_]
  File "/opt/env/haines/dataproc/lib/python2.6/site-packages/pydap/proxy.py", line 115, in __getitem__
    resp, data = request(url)
  File "/opt/env/haines/dataproc/lib/python2.6/site-packages/pydap/util/http.py", line 50, in request
    raise ServerError(msg)
pydap.exceptions.ServerError: 'Server error 42: "\'ntime\'"'

However, we can still read the data via pydap, but have to know to get time separately, e.g.

>>> d2.wtemp.array[:]
array([ 23.10750008,  23.14500046,  23.38870049, ...,  22.0807991 ,
        22.06220055,  22.07489967], dtype=float32)
>>> d2.time.array[:]
array([1320121566, 1320132726, 1320137406, ..., 1322696548, 1322696908,
       1322697268])

So change the following in creator():

# dimension names use tuple so order of initialization is maintained                                                     
    dim_inits = (
        ('ntime', NC.UNLIMITED),
        ('nlat', 1),
        ('nlon', 1),
        ('nz', 1),
        )

to

# dimension names use tuple so order of initialization is maintained                                                     
    dim_inits = (
        ('time', NC.UNLIMITED),
        ('lat', 1),
        ('lon', 1),
        ('z', 1),
        )

and any ntime to time (or nlat to lat or nlon to lon or nz to z, etc)

# (varName, varType, (dimName1, [dimName2],                                                                        
    var_inits = (
        # coordinate variables                                                                                               
        ('time', NC.INT, ('time',)),
        ('lat', NC.FLOAT, ('lat',)),
        ('lon', NC.FLOAT, ('lon',)),
        ('z',  NC.FLOAT, ('z',)),
        # data variables                                                                                                     
        ('wtemp', NC.FLOAT, ('time',)),
        ('cond', NC.FLOAT, ('time',)),
        ('press', NC.FLOAT, ('time',)),
        # derived variables                                                                                                  
        ('depth', NC.FLOAT, ('time',)),
        ('salin', NC.FLOAT, ('time',)),
        ('density', NC.FLOAT, ('time',)),
        )