Missing subdetectorHitNumbers in tracks after ClonesAndSplitTracksFinder

[jmcarcell]

Check duplicate issues.

• Checked for duplicates

Goal

After running ClonesAndSplitTracksFinder the tracks do not preserve the subdetectorHitNumbers. It seems it's simply never set around here: https://github.com/iLCSoft/MarlinTrkProcessors/blob/master/source/Refitting/src/ClonesAndSplitTracksFinder.cc#L235. Since what this algorithm does with the option mergeSplitTracks set to False (this is the current value) is to pick a subset of existing tracks and create a new collection with them, I think it is unexpected not to keep the subdetectorHitNumbers.

Operating System and Version

Alma Linux 9

compiler

GCC 14

The version of the key4hep stack

Nightlies

Package Version

main

Reproducer

source /cvmfs/sw-nightlies.hsf.org/key4hep/setup.sh
ddsim --compactFile $K4GEO/FCCee/CLD/compact/CLD_o2_v07/CLD_o2_v07.xml -G --numberOfEvents 3 --outputFile sim.edm4hep.root --gun.energy '10*GeV' --gun.particle 'mu-' --gun.distribution uniform --gun.multiplicity 10
git clone https://github.com/key4hep/CLDConfig.git
cd CLDConfig/CLDConfig
k4run CLDReconstruction.py --inputFiles sim.edm4hep.root --trackingOnly

and then run the following in python

from podio.root_io import Reader

reader = Reader("output_REC.edm4hep.root")
events = reader.get("events")
event = events[0]
tracksCT = event.get("SiTracksCT")
tracks = event.get("SiTracks")

print(f"Number of tracksCT: {len(tracksCT)}")
print(f"Number of tracks after ClonesAndSplit: {len(tracks)}")

print(f"Number of subdetector hit numbers CT: {len(tracksCT[0].getSubdetectorHitNumbers())}")
print(f"Number of subdetector hit numbers after ClonesAndSplit: {len(tracks[0].getSubdetectorHitNumbers())}")

This prints in my case:

Number of tracksCT: 9
Number of tracks after ClonesAndSplit: 9
Number of subdetector hit numbers CT: 12
Number of subdetector hit numbers ClonesAndSplit: 0

One can do additional checking to see that the tracks are exactly the same except for the subdetector hit numbers.

Additional context

No response

[Zehvogel]

Thanks for opening this issue! I vaguely remember encountering this before but I forgot about it :(

I will pin it to the known issues for now, but you should also consider opening an issue in the affected repository (or even a PR to fix it? :))

[jmcarcell]

There it is: iLCSoft/MarlinTrkProcessors#73