Note that for SparseNDArray it sufficed to have a slotwise int shape: indexing is only by int64 and we took the min to be 0, leaving only the max to be specified. Here, for DataFrame, indices can be of various types so we truly must do parameterizations in terms of domains -- min-max pairs on each dim slot.
Context: https://github.com/single-cell-data/TileDB-SOMA/issues/961
This is a prerequisite for https://github.com/single-cell-data/TileDB-SOMA/pull/1027.
Note that for
SparseNDArray
it sufficed to have a slotwise intshape
: indexing is only byint64
and we took the min to be 0, leaving only the max to be specified. Here, forDataFrame
, indices can be of various types so we truly must do parameterizations in terms of domains -- min-max pairs on each dim slot.