Open imadcat opened 9 years ago
there is currently a work around: in pandas before exporting data frame, set Datetime column as index (or one of the index). then rhdf5 will recognize this index as int64 epoch (nano seconds since 1970-01-01).
it's easy to convert this int64 to date time in R.
the function h5DFread using this work around will soon be included after testing.
It's not implemented in the original package. Reading a hdf5 file with python datetime data field will result a warning and that data field been set to NA.