I agree both friction and frictionless cases need to be analyzed.
I disagree with the practice of completely ignoring friction at the first support next to a pump.
WRC 449 suggests that piping on a can spring will deflect up to 10 mm horizontally without friction (approx. clearance between load column and spring housing).
Suggest using a factor of 1.5 on nozzle allowables for a non-running pump, and for friction case, as it is a transient condition, not permanent.