Hi,

On 2 July 2011 17:07, jpaulini <[email protected]> wrote:

> files, but today I was so surprised that now they implemented a
> different data format...
>
> Looking for some information on how to read the new data with python,
> I only got that it uses HDF file format, but no more...
>

HDF is a very frequenty-encountered format, and as such, there are a few
pythonic ways of reading it in (& saving it, I guess). It would appear that
the format is HDF4. I am partial to GDAL through its python bindings, but it
is a bit limited (if they pack in 1D arrays and stuff, GDAL will just ignore
them). pydf <http://pysclint.sourceforge.net/pyhdf/> deals with all that
stuff. Using gdal to read data into a numpy array (consider that all the
products have the same spatial extent) is quite easy. Suppose you have a
file called 3B42.110531.0.6A.HDF, you just then...

from osgeo import gdal
g = gdal.Open('HDF4_SDS:UNKNOWN:"3B42.110531.0.6A.HDF":0')
precip = g.ReadAsArray()
# precip is a 1440x400 array with the precipitation dataset
g = gdal.Open('HDF4_SDS:UNKNOWN:"3B42.110531.0.6A.HDF":1')
err= g.ReadAsArray()
# err is a 1440x400 array with the relative error


Hope that helps
Jose

Reply via email to