Load a Python/pandas data frame from an HDF5 file into R

loadhdf5data.R

#' @param h5File HDF5 file path
#' @param dataset data frame path in the HDF5 file
#' @examples
#' df <- loadhdf5data("/path/to/file.hdf5", "/path/to/dataset")
#' 
loadhdf5data <- function(h5File, dataset) {
  require(h5) # available on CRAN
  
  f <- h5file(h5File)
  nblocks <- h5attr(f[dataset], "nblocks")
  
  data <- do.call(cbind, lapply(seq_len(nblocks)-1, function(i) {
    data <- as.data.frame(f[paste0(dataset, "/block", i, "_values")][])
    colnames(data) <- f[paste0(dataset, "/block", i, "_items")][]
    data
  }))
  
  h5close(f)
  data
}