[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDF #YZQ-382018]: Variable naming convention in netCDF

Hi Mike,

> In the netCDF docs, I see the following:
>
> 2.1.1 Naming Conventions
>
> The names of dimensions, variables and attributes consist of arbitrary
> sequences of alphanumeric characters (as well as underscore '_' and
> hyphen '-'), beginning with a letter or underscore. (However names
> commencing with underscore are reserved for system use.) Case is
> significant in netCDF names.
>
> Could you please expand on what you mean by 'However names commencing
> with underscore are reserved for system use'? If I name a variable
> starting with underscore, will it work as normal?

We reserved names beginning with an underscore for system use, because
we needed a namespace that wouldn't clash with user names.  Up until
now, we have only used this namespace for special attributes,
including "_FillValue", "_Unsigned", and the special virtual
attributes supported by ncdump and ncgen4 for netCDF-4 variable
characteristics, such as "_Fletcher32", "_ChunkSizes", "_Endianness",
"_DeflateLevel", "_Shuffle", "_Storage", and "_NoFill", and the global
attribute "_Format".  In the future we have plans for other special
attributes, including "_Scale" and "_Offset".

We have so far not used any reserved names for variables, but may want
to do so in the future, and might at some point enforce the
restriction by not permitting new names beginning with "_" in order to
prevent future clashes.  However that would only be for creating new
names.  The netCDF library has always allowed reading names containing
arbitrary characters, so if such files were created in the past, they
will always be readable.

Another consideration is whether you want your files to comply with
the CF Conventions.  The current version 1.4 of CF requires:

   Variable, dimension and attribute names should begin with a letter
   and be composed of letters, digits, and underscores. Note that this
   is in conformance with the COARDS conventions, but is more
   restrictive than the netCDF interface which allows use of the
   hyphen character. The netCDF interface also allows leading
   underscores in names, but the NUG states that this is reserved for
   system use.

so strictly speaking, a file containing a variable beginning with an
underscore that does not have special meaning in the library is not
CF-compliant.

If you want to transform names from one namespace to names that are
guaranteed not to clash with names in another namespace, one way to
do it is to just prefix the names with the namespace followed by
underscore, i.e. "varname" gets mapped to "NCAR_varname", for example,
to use an NCAR namespace.  If you don't care about CF compliance, you
could use the ":" character instead, so "varname" would map to
"NCAR:varname", which would display in the CDL produced by ncdump and
parsed by ncgen as "NCAR\:varname".

I hope this helps to clarify why we reserve leading "_" in names ...

--Russ

Russ Rew                                         UCAR Unidata Program
address@hidden                     http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: YZQ-382018
Department: Support netCDF
Priority: Normal
Status: Closed