Datasets object is implicitly bound to one selected dataset in a counter-intuitive way

[curusarn]

Consider following code:

SOME_ID = "e4c77e0e-2bfa-42e2-baa2-33515e5e45ac"
DIFFERENT_ID = "8b88a424-dbd8-4032-8be7-a930a415b9a5"

ds = Datasets()                # creates Datasets() object
ds.info(id=SOME_ID)            # downloads data and returns info for SOME_ID
ds.info(id=DIFFERENT_ID)       # uses cached data and returns info for SOME_ID