Set of tools that help in reading and writing hdf5 files for .net environments
At the neurology department of the Leiden University Medical Centre in the Netherlands we need to convert large medical data files to a format that could easily be used in programs like Matlab, R and Python.
In the example below an object is created with some arrays and other variables The object is written to a file and than read back in a new object.
private class TestClassWithArray
{
public double[] TestDoubles { get; set; }
public string[] TestStrings { get; set; }
public int TestInteger { get; set; }
public double TestDouble { get; set; }
public bool TestBoolean { get; set; }
public string TestString { get; set; }
}
var testClass = new TestClassWithArray() {
TestInteger = 2,
TestDouble = 1.1,
TestBoolean = true,
TestString = "test string",
TestDoubles = new double[] { 1.1, 1.2, -1.1, -1.2 },
TestStrings = new string[] { "one", "two", "three", "four" }
};
int fileId = Hdf5.CreateFile("testFile.H5");
Hdf5.WriteObject(fileId, testClass, "testObject");
TestClassWithArray readObject = new TestClassWithArray();
readObject = Hdf5.ReadObject(fileId, readObject, "testObject");
Hdf5.CloseFile(fileId);
/// <summary>
/// create a matrix and fill it with numbers
/// </summary>
/// <param name="offset"></param>
/// <returns>the matrix </returns>
private static double[,]createDataset(int offset = 0)
{
var dset = new double[10, 5];
for (var i = 0; i < 10; i++)
for (var j = 0; j < 5; j++)
{
double x = i + j * 5 + offset;
dset[i, j] = (j == 0) ? x : x / 10;
}
return dset;
}
// create a list of matrices
dsets = new List<double[,]> {
createDataset(),
createDataset(10),
createDataset(20) };
string filename = Path.Combine(folder, "testChunks.H5");
int fileId = Hdf5.CreateFile(filename);
// create a dataset and append two more datasets to it
using (var chunkedDset = new ChunkedDataset<double>("/test", fileId, dsets.First()))
{
foreach (var ds in dsets.Skip(1))
chunkedDset.AppendDataset(ds);
}
// read rows 9 to 22 of the dataset
ulong begIndex = 8;
ulong endIndex = 21;
var dset = Hdf5.ReadDataset<double>(fileId, "/test", begIndex, endIndex);
Hdf5.CloseFile(fileId);
This is still an early version of the library. If there are any problems please let me know
Another project on github that uses the HDF.Pinvoke library to read and write HDF5 files is the sharpHDF project. I discovered it while I was working on my own library. It has a different approah to writing and reading hdf5 files. You have to create a Hdf5File object and fill it with groups, attributes and datasets. When you close the Hdf5File object it writes the file.