[Scilab-users] HDF5 write data to file

[Lester Anderson]

Hi all,

The Scilab (SOD) format is the way forward since netCDF and HDF5 are very
similar, but have slightly different naming conventions. I have a
Geographic grid with the structure below, but cannot seem to get the output
file dimensions right, i.e. the x and y coordinates in degrees and not
pixel reference.

File structure from ncdisp('filename')

Source:    Test_10-FAA.grd
Format:    netcdf4

Global Attributes:
           Conventions = "COARDS, CF-1.5"
           title       = ""
           history     = "grdsample -R-10/10/-10/10 -I1m @GMTAPI at -000001
-Gc:\Temp\Test_10-FAA.grd -fg --GMT_HISTORY=false"
           description = ""
           GMT_version = "5.4.3 (r19528) [64-bit] [MP]"
           node_offset = 1
Dimensions:
           lon = 1200
           lat = 1200
Variables:
    lon
           Size:       1200
           Dimensions: lon
           Datatype:   double
           Attributes:
                      long_name    = "longitude"
                      units        = "degrees_east"
                      actual_range = -10  10
    lat
           Size:       1200
           Dimensions: lat
           Datatype:   double
           Attributes:
                      long_name    = "latitude"
                      units        = "degrees_north"
                      actual_range = -10  10
    z
           Size:       1200x1200
           Dimensions: lon,lat
           Datatype:   single
           Attributes:
                      long_name    = "z"
                      _FillValue   = NaN
                      actual_range = -104.9617      456.6511

b=h5open('Test_out.grd','w');
h5write(b,'lat',lat);
h5write(b,'lon',lon);
h5write(b,'data',data);
h5close(b)

save('Test_output.sod','lon','lat','data')

Saving the output creates a file composed of the three datasets (lon, lat,
data). Plotting the data in Scilab via Sgrayplot(x,y,z) it looks correct,
oriented properly and axes of longitude and latitude.

Anyone know how to set the correct attributes to get the HDF5 file
recognised as Geographic?

Lester

On Wed, 25 May 2022 at 15:17, Lester Anderson <arctica1963 at gmail.com> wrote:

> Hi all,
>
> I just found a useful way of getting netCDF files into Scilab via HDF5
> (just found this!):
>
> a=h5open('Test_10-FAA.grd');
> h5ls(a)
> //  "lat"                     "dataset"  //  "lon"                    "dataset"  //  "z"                       "dataset"  //  "Conventions"      "attribute"//  "GMT_version"     "attribute"//  "_NCProperties"   "attribute"//  "description"        "attribute"//  "history"              "attribute"//  "node_offset"       "attribute"//  "title"                   "attribute"
> lon=h5read(a, 'lon');lat=h5read(a, 'lat');data=h5read(a, 'z');
>
> h5close(a);
>
> The data plots fine via Sgrayplot(lon,lat,data,zminmax=[-25,25]) . I need
> to know how to write data out to essentially replicate the input netCDF
> file, so any pointers/examples would be helpful. It would be good to output
> a file in a format that GMT (Generic Mapping Tools) would recognise.
>
> The use of HDF5 seems to solve the issues of working with netCDF-4 files.
>
> Lester
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20220527/96cb243c/attachment.htm>