Explode step filters wrongly?

Created by: tikk3r

Trying to split out multiple directions and filtering stations using the explode step seems to do something strange with the filtering. Running

msin.ntimes=200    
msout.storagemanager = dysco    
    
steps = [explode]    
explode.steps = [shift, adder, filter, avg2, msout]    
explode.replaceparms = [shift.phasecenter, msout.name]    
    
shift.type = phaseshift    
shift.phasecenter=[[163.026994deg,57.685891deg],[162.839619deg,58.446268deg]] 
    
adder.type = stationadder    
adder.stations = {ST001:'CS*'}    
    
filter.type = filter    
filter.baseline=!CS*&&*    
filter.remove = True    
    
avg2.type = average    
avg2.timestep = 2    
avg2.freqstep = 2    
    
msout.overwrite = True    
msout.name = [P3.ms, P42.ms] 

gives me a measurement set with only CS stations present, whereas running

msin.ntimes=200

msout=P3.superstation.ms
msout.storagemanager=dysco

steps=[shift,adder,filter,averager]

shift.type=phaseshift
shift.phasecenter=[163.026994deg,57.685891deg]

adder.type=stationadder
adder.stations={ST001:'CS*'}

filter.type=filter
filter.baseline=!CS*&&*
filter.remove=True

averager.type=averager
averager.freqstep=2
averager.timestep=2

gives me the expected output: a MS with only remote and international stations and the ST001 station.