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

[netCDF #LZQ-123544]: Curvlinear Coordinates dimension sizes

Hi Aaron,

There are several ways this must be considered, in terms of whether or not it 
is valid.  First, there are conventions to consider; does this adhere to the CF 
conventions? This question likely doesn't matter, as you don't reference any 
particular conventions.  If I understand correctly, you are just curious if 
this is *technically* valid, e.g. will the netCDF library support this.

Coordinate variables are single-dimension in nature, as described here:  

* 
http://www.unidata.ucar.edu/software/netcdf/docs/BestPractices.html#bp_Coordinate-Systems

For multi-dimensional "coordinate variables", the nomenclature would be an 
"auxiliary" coordinate variable, described here:

* 
http://cfconventions.org/cf-conventions/v1.6.0/cf-conventions.html#coordinate-system

You'll note this is part of the CF conventions and not *yet* in the netCDF best 
practices guide; I'm working on fixing that, however.  The safe thing about 
using these guidelines is that they conform to how the code is written, and 
deviating from these guidelines can result in (currently) undefined and 
unexpected behavior under specific circumstances; see 
https://github.com/Unidata/netcdf-c/issues/295 for a relevant example I have 
been actively working on for the last few weeks.

This is all separate from the question, however, of what does VisIt support?  
Unfortunately, I am not familiar with VisIt, and you may need to contact their 
support for specifics as to how they handle this.

I hope this information is useful,

-Ward






> Hi,
> 
> I have a question about the way that coordinate variables work with
> NETCDF4. The code I am running solves the Navier-Stokes equations in a box
> which can have the z dimension mapped as a function of x, so as to add
> topography. I have managed to get it to output into NETCDF for both cases;
> with and without topography. In the code I define the dimensions to be
> their respective number of points and then during the nc_put_vara_double
> command I put each dimension as a 3D array. However when I attempt to view
> the data in VisIt the mapping does not appear and the domain still looks
> like a box  but with the data warped inside. I have opened the file in
> MATLAB and I can easily extract the information and get the topography. I
> am not sure if this is a VisIt problem or a NETCDF one. Specifically is it
> valid to assign the coordinate variable to a 3D array?
> 
> Thanks,
> -Aaron Coutino
> 
> 


Ticket Details
===================
Ticket ID: LZQ-123544
Department: Support netCDF
Priority: Normal
Status: Closed
===================
NOTE: All email exchanges with Unidata User Support are recorded in the Unidata 
inquiry tracking system and then made publicly available through the web.  If 
you do not want to have your interactions made available in this way, you must 
let us know in each email you send to us.