On 11/16/10 8:44 AM, Gerrit Holl wrote:
> [Gerrit Holl]
>>> I am moving from NetCDF to HDF5. In NetCDF, I can assign attributes to
>>> each variable/column. That is very useful for self-documenting the
>>> file like this:
>>>
>>> double AVHRR_LONG(Collocations) ;
>>> AVHRR_LONG:long_name = "AVHRR Longitude" ;
>>> AVHRR_LONG:units = "degrees_east" ;
>>> AVHRR_LONG:valid_range = -180., 180. ;
>>>
>>> I know that in HDF5, I can assign attributes to each node. But I think
>>> the columns of a table are not nodes. Is there an equivalent in HDF5
>>> to NetCDF's variable attributes? How would I use this in pytables?
> [Jeff Whitaker]
>> Gerrit: Are you aware that if you use netcdf 4, then your data is
>> actually saved in an HDF5 file (readable by hdf5 and netcdf clients)?
> A bit, but I don't understand the details of it. I should specify that
> I was using NetCDF 3 before.
>
> But then how would I use all of pytables' nice HDF5 features (blosc
> compression, fast searching, possibly indexing if I buy the
> professional version) in a NetCDF4 file? Some combination of your
> netcdf4-python with pytables?
Gerrit: You can use zlib compression (with the shuffle filter) with
netcdf, but not those other features you mention. It's possible that a
netcdf 4 compatible HDF file can be opened with pytables, but I've
never tried that. The main benefit of sticking to the netcdf api for
creating HDF5 is interoperability - netcdf is pretty much the de-facto
standard in the geophysical sciences.
-Jeff
> [David E. Sallis]
>> You can store arbitrary key/value pairs on a node by using the node's
>> _f_setAttr() method. I use this frequently and in much the same way as in
>> your
>> example. The key is the column name, and the value is a Python dictionary
>> containing the long name, units, etc.
>>
>> Example: n._f_setAttr('AVHRR_LONG':{'long_name':'AVHRR_Longitude',
>> 'units':'degrees_east', 'valid_range':(-180.0,180.0)})
> What is 'n' in this case? The table containing those columns? So then
> I actually set the properties to the table, with property names equal
> to the column, rather than setting properties to the variable, or am I
> misunderstanding things?
>
> Gerrit.
>
--
Jeffrey S. Whitaker Phone : (303)497-6313
Meteorologist FAX : (303)497-6449
NOAA/OAR/PSD R/PSD1 Email : jeffrey.s.whita...@noaa.gov
325 Broadway Office : Skaggs Research Cntr 1D-113
Boulder, CO, USA 80303-3328 Web : http://tinyurl.com/5telg
------------------------------------------------------------------------------
Beautiful is writing same markup. Internet Explorer 9 supports
standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3.
Spend less time writing and rewriting code and more time creating great
experiences on the web. Be a part of the beta today
http://p.sf.net/sfu/msIE9-sfdev2dev
_______________________________________________
Pytables-users mailing list
Pytables-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/pytables-users
