Re: [netcdfgroup] netcdf4 -problems

• To: Henry Butowsky <henryb@xxxxxxxxxxxx>
• Subject: Re: [netcdfgroup] netcdf4 -problems
• From: Ed Hartnett <ed@xxxxxxxxxxxxxxxx>
• Date: Fri, 11 Jan 2008 08:28:22 -0700

Henry Butowsky <henryb@xxxxxxxxxxxx> writes:

> Hi all,
>
> It seems that using netcdf4 API  with hdf5 files I have the following problem:
>
> If a dimension is defined in the output file  -- then a  variable of the same 
> name as the dimension  defined in output MUST have a single dim which is the 
> same
> name a the defined dimension. This is not the case with netcdf3.
> what I am saying is that the defined var  is OK as a co-ordinate var but not 
> anything else
>
> The below script illustrates the problem:
>
> defdim("a",2);             //  dim "a=2" defined in output             
> defdim("b",3);             //  dim "b=3" defined in output
> defdim("c",4);             //  dim "c=3"  defined in output
> a[$a,$b]=10.0;           // variable "a" defined as "double a[a,b];"  works 
> in netcdf3 but NOT netcdf4         
> c[$c]=20.0                  // variable "c" defined as "double c[c];"    
> works in metcdf3 and netcdf4     
>
> Regards Henry
>

This is a netCDF-4 bug.

Thanks for pointing this out Henry, I will get this fixed before the
next beta release.

Thanks again,

Ed

-- 
Ed Hartnett  -- ed@xxxxxxxxxxxxxxxx

• References:
  • [netcdfgroup] netcdf4 -problems
    • From: Henry Butowsky