[Python] Writing complex Pandas DataFrame to HDF5 using h5py

I have a Pandas DataFrame with mixed scalar and array-like data of different raw types (int, float, str). The DataFrame's types look like this:

'col1', dtype('float64')
'col2', dtype('O') <-- array, item type str
'col3', dtype('int64'
'col4', dtype('bool')
...
'colA', dtype('O') <-- array, item type str
...
'colB', dtype('O') <-- array, item type float
...
'colC', dtype('O') <-- scalar, str
...
'colD', dtype('O') <-- array, item type int
...
some more mixed data type columns


The length of the numeric array-like data is variable fron DataFrame row to row.

Currently, I try to write this DataFrame naively to a HDF5 file by

with h5py.File(self.file_path, 'w') as f:
f.create_dataset(dataset_name, data=dataframe)


This operation fails with the error message

tid = h5t.py_create(dtype, logical=1)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "h5py/h5t.pyx", line 1669, in h5py.h5t.py_create
File "h5py/h5t.pyx", line 1693, in h5py.h5t.py_create
File "h5py/h5t.pyx", line 1753, in h5py.h5t.py_create
TypeError: Object dtype dtype('O') has no native HDF5 equivalent


What is a way / the best way to write this data to HDF5?

Continue reading...