See title. For example, splitting a data object of shape (972, 1) at one value along the 972 length axis generates a collection of two data objects of lengths (685,) and (287,), where intended behavior would be (685,1) and (287,1). In methods meant to allow for different slice widths, e.g. also (972, 3), or more generally written for 2D data, losing the original dimensionality for only slices of width 1 requires special case processing.

See title. For example, splitting a data object of shape (972, 1) at one value along the 972 length axis generates a collection of two data objects of lengths (685,) and (287,), where intended behavior would be (685,1) and (287,1). In methods meant to allow for different slice widths, e.g. also (972, 3), or more generally written for 2D data, losing the original dimensionality for only slices of width 1 requires special case processing.