NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
Hi Harvey,
On Thu, 26 Apr 2001, Davies, Harvey wrote:
> > A dataset that was written with a _FillValue within a specified valid
> > range
> > would not have that _FillValue treated as the valid default data value by
> > many existing applications, i.e., such a dataset would break most existing
> > applications.
> > :
>
> I do not understand how this would break most existing applications.
My experience with generic applications is that if values indicated by the
_FillValue attribute are treated in some special way, they are treated that
way independently of whether or not the _FillValue falls within a specified
valid range.
This is true for the first two generic applications that I tested, namely
ncdump and ncview.
ncdump uses `_' to represent data values that are equal to
the `_FillValue' attribute for a variable, intended to
represent data that has not yet been written.
I verified that this behavior is true even when the _FillValue is within
the valid range. This is not broken in the sense that ncdump is doing
exactly what it says it does. But I think that most people would be
unhappy to see '_' in their ncdump output for values that they consider to
be valid.
ncview also treats _FillValue as missing even if it's contained in a
specified valid range. In an image of a 2D field this results in "valid"
data being displayed in a color outside the color map.
My conclusion from this is that _FillValue should never be used to indicate
a default value for valid data, because it will be treated exactly as if it
were missing, or never written, by most applications.
Brian
-----------------------------------------------------------------
Brian Eaton | email: eaton@xxxxxxxx
Climate Modeling Section |
National Center for Atmospheric Research |
P.O. Box 3000, Boulder, CO 80307 |
netcdfgroup archives: