Thanks, Russ. This is a fine explanation, but it would help to have
the on-line documentation a bit more explicit about this. Perhaps a
note along the lines of your paragraph 2 below would be good at the
top of the error code pages. The tip about using nc_strerror for
unknown codes is especially helpful.
http://www.unidata.ucar.edu/software/netcdf/docs/netcdf-c/Error-Codes.html
It would also be helpful to have error code pages in the Fortran docs.
--Dave
Russ Rew wrote:
Hi Nate,
You wrote:
The error code 24 is a POSIX error code that is happening at a lower
level than NetCDF, that is all I can say, i.e. the operating system is
generating an error not NetCDF.
Right, the netCDF C library returns negative error codes for
netCDF-specific errors and positive error codes (corresponding to what's
defined in /usr/include/errno.h) for system-level errors. If you call
the nc_strerror() function (or NF_STRERROR for F77 or NF90_STRERROR for
F90) with the error code as argument, it returns an error message. This
works for either kind of error, so you don't have to know whether it's
netCDF-specific or systems-level.
The reason there are both netCDF-specific and systems-level errors for
too many open files is historical. An earlier implementation of the
library put limits on the number of open netCDF files that were more
restrictive than the operating system's open file limit, but that is no
longer the case.
--Russ
On 10/23/07, Dave Allured <dave.allured@xxxxxxxx> wrote:
All,
According to the docs for the Netcdf C interface, Nate's error code
for too many files open should be -34, not 24. 24 is not listed.
http://www.unidata.ucar.edu/software/netcdf/docs/netcdf-c/Error-Codes.html
Does anyone know the reason for this discrepancy? I ask for
reference only. Thanks.
--Dave A.
CU/CIRES Climate Diagnostics Center (CDC)
NOAA/ESRL/PSD, Climate Analysis Branch (CAB)
http://www.cdc.noaa.gov
Nathaniel Nutter wrote:
OK, for those who encounter this later it is because despite the files
being open read only they do need to be closed. Thanks to: giuliani,
Hannes, and Timothy for point this out to me so quickly. :-)
I'm brand new at this stuff so I appreciate the help immensely!
On 10/22/07, Nathaniel Nutter <iam@xxxxxxxxxxx> wrote:
I'm having a hard time searching for this on the mailing list archive
and Google due to the unhelpfully constraining nature of a number and
'return'.
I am running a program that reads in data every 'month' and it goes
through many years and then mysteriously returns non-zero after
opening the same files dozens of times.
Every time prior (again dozens) it returns zero. The file is being
opened read-only, no changes are being made to it.
Debugging info from my program:
nc_open status = 24
filename = dat/ts.nc
Nate Nutter
iam@xxxxxxxxxxx
_______________________________________________
netcdfgroup mailing list
netcdfgroup@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit: http://www.unidata.ucar.e
du/mailing_lists/
_______________________________________________
netcdfgroup mailing list
netcdfgroup@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit: http://www.unidata.ucar.edu
/mailing_lists/
_______________________________________________
netcdfgroup mailing list
netcdfgroup@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit: http://www.unidata.ucar.edu/m
ailing_lists/
