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

Re: examples

>From:  "Reula, Oscar" <address@hidden>
>Subject: Re: 20020910: subject
>Organization: FaMAF, Univ. Nac. de Cordoba, 5000 Cordoba, ARGENTINA
>Keywords: 200210031824.g93IO1129487, no data written

> Russ, I printout the start/count values, they are OK, and the size
> of the files grow, so something is going there.. mistery!
> 
> Shouldn't ncdump see those values?
 ...
> writing data in netcdf files
> start_field = 0, 0, 0, 0 count_field = 0, 26, 26, 11, index_time = 0
> finishing writing data in netcdf files

This would not have written anything, because count_field[0] is 0.
The count_field is the number of slices along a dimension that are
written, so the total number of values written should be the product
of the count_field elements, 0*26*26*11.

> writing data in netcdf files
> start_field = 1, 0, 0, 0 count_field = 1, 26, 26, 11, index_time = 1
> finishing writing data in netcdf files

This should have written 1*26*26*11 values.

> rwriting data in netcdf files
> start_field = 2, 0, 0, 0 count_field = 2, 26, 26, 11, index_time = 2
> finishing writing data in netcdf files

This should have written 2*26*26*11 values.

I think you probably just want to write 1*26*26*11 values in each
case, so I don't think you should be incrementing count_field[0] and
it should be initialized as 1 before the loop.

Nevertheless, ncdump should see the values you've written once the
file is closed.  It may or may not see them prior to this, depending
on the system you are using and when it synchronizes memory buffers of
an open file to disk.  If you call nc_sync() or nc_close(), an effort
is made to synchronize everything to disk, so ncdump should see it
then, on a local file system.  On the other hand if you are writing to
a remotely mounted file system, the synchronization semantics may be a
little less clear.  Are you saying that later, even after the file is
closed, ncdump doesn't show the data has been written?

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden                     http://www.unidata.ucar.edu