Jed,
I don't know of any way to read an arbitrary 2-D mapping of a 1-D,
netCDF variable into memory with a single call to the C or Fortran
interface.
You might be able to do this with the Java interface, however.
Regards,
Steve Emmerson
>Date: Thu, 26 Feb 2004 18:00:46 +0100
>From: "Jed O. Kaplan" <jed.kaplan@xxxxxx>
>To: netcdfgroup@xxxxxxxxxxxxxxxx
>Subject: reading a 1D variable into 2D memory
>
> I have a netCDF file with data stored in a one-dimensional variable
> that represents two-dimensional data in physical representation i. e.
> the 2-D data has been collapsed into 1-D.
>
> I would like to read in a rectangular subset (hyperslab) of this
> variable into a 2-D array in a program. I can calculate the starting
> point for the read and the number of values to be read, but in this
> case, the hyperslab of values are not contiguous in the variable in the
> netCDF file. I want to read e. g. 3 values, then skip 100 values, then
> 3 more values, skip 100, etc.
>
> At the risk of having missed something really obvious, is there a way
> to read the values without making multiple calls to nf90_get_var()? The
> "stride" argument to get_var seems to only allow me to read one value
> before skipping ahead in the file; what do I do to read a series of
> values?
>
> I am using netCDF for Fortran90 v 3.5 on MacOS X 10.3
>
> Thank you for your help.
>
> Cheers,
>
> Jed Kaplan
