[gradsusr] GrADS Control File for Array with Spectral Band Dimension

[Jennifer Adams]

Hi, James -- 
I am not able to reproduce this problem. I can display all the variables using your full descriptor on my OSX laptop and our CentOS servers. Perhaps there is a memory issue -- are you working on a platform that has little available memory? 
--Jennifer



On Sep 20, 2013, at 10:15 AM, James E. Johnson wrote:

> On Thursday 19 September 2013 06:11:44 pm James E. Johnson wrote:
>> On Thursday 19 September 2013 04:55:58 pm Jennifer Adams wrote:
>>> First of all, thank you for providing all the information and data I
>>> needed to reproduce the error! The failure of method 1 (using array
>>> indices in the units field of the variable declaration instead of
>>> x,y,z,t,e) is indeed a bug that is unique to the hdf5 data type. The
>>> critical bit of information in the error message from the HDF5 library
>>> was this: "src and dest data spaces have different sizes." In method 1,
>>> the variables are described to GrADS as being only 2-dimensional (i.e.,
>>> not Z-varying) even though they really have 3 varying dimensions, and the
>>> library trips on this deception. The netcdf library is not as fussy. I
>>> will put a fix in version 2.1, but for now you will have to use method 2.
>>> --Jennifer
>> 
>> Thanks Jennifer,
>> 
>> I think there may be another bug. I have a more extensive control file with
>> all the variables from the file. I can successfully display data from the
>> first variable I read, but subsequent variables complain:
>> 
>> HDF5-DIAG: Error detected in HDF5 (1.8.7) thread 0:
>>  #000: H5Shyper.c line 6593 in H5Sselect_hyperslab(): not a data space
>>    major: Invalid arguments to routine
>>    minor: Inappropriate type
>> HDF5 Error: unable to select dataspace hyperslab
>> Data Request Error:  Error for variable 'uvai'
>>  Error ocurred at column 1
>> DISPLAY error:  Invalid expression
>>  Expression = uvai
> 
> One thing I notice is that if I close the control file, and reopen it, then I 
> can read another variable. But it seems GrADS 2.0.2 has a problem somewhere 
> allowing you to read/displaying mutliple variables.
> 
> -----------------------------------------------------------------------------
> | James E. Johnson                   | address:                              |
> | ADNET Systems, Inc.                |  NASA Goddard Space Flight Center     |
> | e-mail: james.johnson at nasa.gov     |  Data and Information Services Center |
> | phone:  301-614-5121               |  Code 610.2, Bldg 32, Room S130G      |
> | fax:    301-614-5268               |  Greenbelt, MD  20771                 |
> ------------------------------------------------------------------------------
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr

--
Jennifer M. Adams
Center for Ocean-Land-Atmosphere Studies (COLA)
111 Research Hall, Mail Stop 2B3
George Mason University
4400 University Drive
Fairfax, VA 22030 





-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20130920/e41ce57a/attachment.html