mistake in HDF 3.0 file spec

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

mistake in HDF 3.0 file spec

Krug, Markus

Dear all,

 

I had a closer look to the HDF 3.0 file spec and especially on the data type messages (https://support.hdfgroup.org/HDF5/doc/H5.format.html#DatatypeMessage). As far as I understand is the field ‘Dimension #n Size’ for the compound datatype version 1 defined within 4 bytes. From analysis of a given hdf-file I think it must be a 8-byte value.

Can someone confirm that or let me know why I’m wrong?

 

Best Regards

Markus

 


_______________________________________________
Hdf-forum is for HDF software users discussion.
[hidden email]
http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5
Reply | Threaded
Open this post in threaded view
|

Re: mistake in HDF 3.0 file spec

Quincey Koziol-3
Hi Markus,
The dimension size for array datatypes is limited to 32-bits, but the size for dataspace dimensions is 64-bits.

Quincey


On Aug 25, 2017, at 4:45 AM, Krug, Markus <[hidden email]> wrote:

Dear all,
 
I had a closer look to the HDF 3.0 file spec and especially on the data type messages (https://support.hdfgroup.org/HDF5/doc/H5.format.html#DatatypeMessage). As far as I understand is the field ‘Dimension #n Size’ for the compound datatype version 1 defined within 4 bytes. From analysis of a given hdf-file I think it must be a 8-byte value.
Can someone confirm that or let me know why I’m wrong?
 
Best Regards
Markus
 
_______________________________________________
Hdf-forum is for HDF software users discussion.
[hidden email]
http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5


_______________________________________________
Hdf-forum is for HDF software users discussion.
[hidden email]
http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5