I will investigate.
=Dennis Heimbigner
Unidata
ed zaron wrote:
Hi James and Dennis,
I discovered an apparent bug in the latest beta snapshot concerning nccopy.
I have two builds of netcdf, and I executed the following commands and obtained
different results:
USING LATEST BUILD (latest beta referenced in previous email)
ezaron$ ~/opt/bin/nccopy
http://omglnx1.meas.ncsu.edu:8080/thredds/dodsC/gomexppp/2010_June/avg/N N.nc
USING AN OLDER BUILD:
ezaron$ /opt/local/bin/nccopy
http://omglnx1.meas.ncsu.edu:8080/thredds/dodsC/gomexppp/2010_June/avg/N N.nc
The latest build fails to copy all the variables correctly. For some reason the
file it creates discards all the time variability of the 3-dimensional fields
(u,v, temp, and salt). I have attached the output of ncdump -h for the files. I
can post the entire files if needed, they are about 1GB.
Please see below the nc-config information for both netcdf builds.
Here's the config info on the LATEST BUILD:
ezaron$ ~/opt/bin/nc-config --all
This netCDF 4.1.2-beta2-snapshot2011011022 has been built with the following features:
--cc -> cc
--cflags -> -I/Users/ezaron/opt/include -I/opt/local/include
-I/opt/local/include -I/opt/local/include
--libs -> -L/Users/ezaron/opt/lib -lnetcdf -L/opt/local/lib -lhdf5_hl
-lhdf5 -L/opt/local/lib -lz -L/opt/local/lib -lcurl -L/opt/local/lib
-L/opt/local/lib -L/opt/local/lib -lidn -lssl -lcrypto -lssl -lcrypto -lz -lz -lz
--cxx -> g++
--has-c++ -> yes
--fc -> g95
--fflags -> -g -O2 /Users/ezaron/opt/include
--flibs -> -L/Users/ezaron/opt/lib -lnetcdff -lnetcdf
--has-f77 -> no
--has-f90 -> no
--has-dap -> yes
--has-nc2 -> yes
--has-nc4 -> yes
--has-hdf5 -> yes
--has-hdf4 -> no
--has-pnetcdf-> no
--has-szlib -> no
--prefix -> /Users/ezaron/opt
--includedir-> /Users/ezaron/opt/include
--version -> netCDF 4.1.2-beta2-snapshot2011011022
Here's the config info on the working OLDER BUILD:
ezaron$ /opt/local/bin/nc-config --all
This netCDF 4.1.1 has been built with the following features:
--cc -> /opt/local/bin/gcc-mp-4.3
--cflags -> -I/opt/local/include
--libs -> -L/opt/local/lib -lnetcdf
--cxx -> /opt/local/bin/g++-mp-4.3
--has-c++ -> yes
--fc -> /opt/local/bin/gfortran-mp-4.3
--fflags -> -O2 -m64 -I/opt/local/include
--flibs -> -L/opt/local/lib -lnetcdff -lnetcdf
--has-f77 -> yes
--has-f90 -> yes
--has-dap -> yes
--has-nc2 -> yes
--has-nc4 -> yes
--has-hdf5 -> yes
--has-hdf4 -> no
--has-szlib -> yes
--prefix -> /opt/local
--includedir-> /opt/local/include
--version -> netCDF 4.1.1
Please forgive me if this is not the right place to post a bug report.
All the best,
-Ed
------------------------------------------------------------------------
------------------------------------------------------------------------
On Jan 10, 2011, at 4:21 PM, Dennis Heimbigner wrote:
The netcdf snapshot
(ftp://ftp.unidata.ucar.edu/pub/netcdf/snapshot/netcdf-4-daily.tar.gz)
available TOMORROW (Tuesday)
now supports the ability to access opendap data stored in files.
If you have a file called, say, dataset.dods, it is expected that
it contains the on-the-wire data for a request to an opendap server.
Such a file can be created using, for example, wget or a web browser.
Similarly, one can save .das and .dds data in files.
Suppose you have such a file at absolute path /home/abc/dapdata.dods.
You should now be able to access it using, for example the command:
ncdump file:///home/abs/dapdata
You could also use nccopy instead of ncdump.
NOTE: the final .dods is left off.
There are some things to note:
1. You must have at least a .dods file (e.g. dataset.dods).
2. If there is also a /home/abs/dapdata.das,
it will be used to obtain attributes.
3. If there is a /home/abs/dapdata.dds
it will be used. If this does not exist,
then the dds part of the .dods file will be used.
If you are in a position to try this out, please do so
and let me know if there are any problems.
=Dennis Heimbigner (dmh@xxxxxxxx)
Unidata
