Re: [netcdf-java] related to HDF-5 shared dimension

On 3/7/2012 7:53 PM, ghansham sangar wrote:
R/Sir

With the reference to this link:

http://www.unidata.ucar.edu/mailing_lists/archives/cf-satellite/2011/msg00069.html

I have pasted the contents of the link here


  Re: [cf-satellite] Query related to CF conventions

  * /To/: cf-satellite@xxxxxxxxxxxxxxxx
    <mailto:cf-satellite@DOMAIN.HIDDEN>
  * /Subject/: Re: [cf-satellite] Query related to CF conventions
  * /From/: John Caron <caron@xxxxxxxxxxxxxxxx
    <mailto:caron@DOMAIN.HIDDEN>>
  * /Date/: Fri, 29 Jul 2011 06:40:45 -0600

On 7/28/2011 10:32 PM, ghansham sangar wrote:

    R/Sir


    I have an issue related to CF conventions 2D coordinate variables
    (I mean lat/lon arrays corresponding to satellite images) I have
    been able to handle the case where lat/lon arrays have one to one
    correspondence with satellite imagery pixels.

    But if we have some scaling like:
    1. Image Data ( 5600* 5600)
    Lat/Lon Arrays (1400*1400) each

    Or I have the other ways
    2. Image Data(1400x1400)
    Lat/Lon Arrays (2800*2800).


    How to indicate that in the image pixel should refer to which
    Lat/Lon array pixel. Like in first case every four pixels of image
    correspond to 1 lat/lon array pixel.

    I am using HDF 4 for writing products.

    I have tried to search on internet. It says something about
    dimscale. But I am not aware what exactly to do.

    regards
    Ghansham


    _______________________________________________
    cf-satellite mailing list
    cf-satellite@xxxxxxxxxxxxxxxx
    For list information or to unsubscribe, visit:
    http://www.unidata.ucar.edu/mailing_lists/

HI Ghansham:

1. I understand that HDF-EOS has a convention for your situation. Are you using that? Could you explain how it works?2. CF uses netCDF as its file format. HDF4 lacks shared dimensions, so is not an complete fit to the CF model. Why are you writing in HDF4?
John




I  am not using HDF-EOS.
I am sorry in the last mail, I mentioned it was HDF-4. But as in the coming time we plan to generate most of the data products in HDF-5.
Is it now possible in HDF-5 context?
HDF-5 has dimension scale feature.
I also read that HDF-5 dimension scale is not exactly what is netcdf shared dimension. But now I want to know is there a way in which we can accommodate such datasets in HDF-5. I mean to say can we have a 2-D coordinate variable (as stated in the previous mail) in HDF-5:

Image Data (5600*5600)
Lat Array (1400*1400), Lon Array (1400*1400)

OR
Image Data (1400*1400)
Lat Array (2800*2800), Lon Array (2800*2800)
regards
Ghansham



HDF5 has dimension scales, but they are not the same as shared dimensions (close, no cigar).

You want to use netCDF4 for this. It writes HDF5 files, with a kludge, I mean workaround, for shared dimensions.

John
  • 2012 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdf-java archives: