tatami-inc/tatami_hdf5

Fix the various TODOs in the HDF5 sparse reader.

LTLA opened this issue · 0 comments

LTLA commented
  • Use the dataset chunk size in the fapl.
  • Avoid loading all data for sliced queries. (Probably take the range of indices to simplify the code and avoid freaking out HDF5 with hyperslab unions for individual observations.)