[CF-metadata] missing_value vs. _FillValue

From: Brian Eaton <eaton>
Date: Thu, 6 Nov 2003 10:11:49 -0700

Hi Russ,

On Wed, Nov 05, 2003 at 10:40:28PM -0700, Russ Rew wrote:
> The NUG (NetCDF User's Guide) does not deprecate the missing_value
> attribute, but instead states that its purpose is different from the
> purpose for _FillValue:

Thanks for your response.

I agree that the NUG doesn't deprecate missing_value. (I apologize for
misstating this in the current CF doc.) But the statement "This attribute
is not treated in any special way by the library or conforming generic
applications" is about as close to deprecation as you can get without doing
so explicitly :-)

> valid_range suitable for use as a missing value. You may be right
> about Harvey deprecating it in an email message, but I think we can
> use the NUG definition above as providing a useful meaning for it.

I don't think the missing_value as currently defined in the NUG is useful.
The main problem is that its actual use is quite different from what a user
would expect, and hence its use will be highly error prone. One would
expect that by assigning values to the missing_value attribute that a
generic application would recognize those values as missing. Not true.
According to the NUG, values assigned to the missing_value attribute will
not be recognized as missing by generic applications unless those values
are outside the valid range. The default valid range, set by the default
_FillValue, is not easily determined. So in practice it seems reasonable
to say that the valid range must be explicitly set (via setting the
appropriate valid_range, valid_min, valid_max, or _FillValue attributes) in
addition to setting the missing_value attribute. The really unfortunate
part is that even if the data writer gets it right, based on my experience
a generic application probably won't recognize values as missing unless the
value exactly matches the _FillValue.

Brian
Received on Thu Nov 06 2003 - 10:11:49 GMT