change of status of particle ("outside") not in output netcdf file

[Sandra-Esther]

Hi,
I recently switched from a quite old version to a new version of Opendrift (v1.14.3), and while running my tests with the new version, I noticed that when I write out the trajectories into a netcdf file, the change in status in case a particle gets deactivated when leaving the domain ("outside") is not written into the netcdf file but remains zero, before it changes to NaN. I used e.g. "o.set_config('drift:deactivate_north_of', 68)" to deactivate the particles. The netcdf header is fine and it works in case a particle is deactivated when stranding.
I made a little test run from a general example in the Opendrift folder, which is attached (python script in txt file), and would like to ask if someone could please have a look and, if I am not wrong, maybe have an idea how to fix it. (I could not attach the resulting netcdf files)
Thank you very much for your help!
test_deactivate.txt

[knutfrode]

Hi,
I simplified your example a little more, and yes you are right - the "outside" status values do not seem to be written to file - but are available in the data after simulation.
I will investigate this.

import opendrift
from datetime import datetime, timedelta
from opendrift.readers import reader_oscillating
from opendrift.models.oceandrift import OceanDrift

# Oscillating east-west current to illustrate stranding options
reader_osc = reader_oscillating.Reader('x_sea_water_velocity', amplitude=1,
                                       zero_time=datetime.utcnow())

o = OceanDrift(loglevel=20)
o.add_reader([reader_osc])
o.set_config('environment:fallback:y_sea_water_velocity', .2)
o.set_config('drift:deactivate_east_of', 13.0)
o.set_config('general:coastline_action', 'stranding')
o.seed_elements(lon=12.2, lat=67.7, radius=5000, number=500, time=reader_osc.zero_time)
outfile = 'test_status_deactivate.nc'
o.run(duration=timedelta(hours=24), outfile=outfile)

o.plot()  # We see the deactivated elements

o = opendrift.open(outfile)
o.plot()  # After reimport we do not see the deactivated elements

[knutfrode]

It appears that this commit fixed this issue: 43764a9

So please update with git pull to see if this works also for you.
Note that this update is not yet on Conda.

[Sandra-Esther]

It now works fine for me, with the simple test example as well as my more complex setup.
Thank you very much!