I have been noticing some inconsistency with the behavior of shock_spectrum() with aggregate_axis = True. When the dataset is short, shock_spectrum outputs only the “Resultant” column. When the dataset is longer, shock_spectrum outputs the “X”, “Y”, “Z”, and “Resultant” columns.
I believe the inconsistency was introduced in PR-191 (#191).
The intended behavior should be documented and the shock_spectrum() and rolling_shock_spectrum() corrected to match.