h5readAttribute#

Purpose#

Read attributes from an HDF5 file into GAUSS.

Format#

attr_read = h5readAttribute(fname, dname, attr_name)#

Parameters:

fname (string) – a name of HDF5 file.
dname (string) – a name of the HDF5 dataset.
attr_name (string) – the name of attribute.

Returns:

attr_read (matrix, array or string array) – entire contents of the HDF5 dataset attribute.

Examples#

Create an HDF5 dataset and add headers#

// Define filename
fname = "commodities.h5";

// Define dataset within file
dname = "/energy";

// Define a size of 100 rows and 4 columns
r = 100;
c = 5;
dims  = r|c;

// Create an HDF5 dataset with room for 100 observations of 4 variables
call h5create(fname, dname, dims);

// Variable names for the dataset
attr = "Crude Oil"$|"Gasoline"$|"Heating Oil"$|"Diesel";

// Define a name of the attributes
attr_name = "headers";

// Write attributes to a HDF5 file
call h5writeAttribute(fname, dname, attr_name, attr);

// Read attributes from a HDF5 file
attr_read = h5readAttribute(fname, dname, attr_name);

Add data and calculate descriptive statistics#

// Set seed for repeatable random data
rndseed 54235;

// Create 100x4 random normal data
x = rndn(100, 4);

// Write data to dataset created in the example above
call h5write(fname, dname, x);

/*
** Calculate descriptive statistics on some of the variables
** using an hdf5 file schema (h5://filename/dataset)
*/
call dstatmt("h5://commodities.h5/energy", "Gasoline + Heating Oil");

----------------------------------------------------------------------------------
Variable          Mean   Std Dev    Variance   Minimum   Maximum   Valid   Missing
----------------------------------------------------------------------------------

Gasoline        0.0212    1.0130      1.0261   -2.9943    2.3527     100      0
Heating Oil    -0.1120    0.9263      0.8580   -2.7726    3.0910     100      0

Remarks#

HDF5 does not support partial read or write of dataset attributes. The entire contents of the attribute will be read.
GAUSS functions that accept HDF5 datasets as a data source, expect the dataset to have an attribute named "headers", containing the variable names of the dataset.