I can confirm this. It appears that when running identical simulations, changing only the output module, reconstruction data is seen in ROOT output but not in JSON. This may follow on from a change I made yesterday, where the I moved the SciFiHit data from from being stored in the SciFiEvents in the recon branch, to being stored in the SciFiHits member of the MC events in the MC branch. Should not have made any difference to the output routines, but should and did are not the same. Will need to investigate to see if this is a fault in the SciFi code or in the Processor framework, which should be doing all the i/o for this.

What are you doing with my lovely MC data structure?

Solved. The json data was there, just for some reason there were lots of empty events at the start (hmmm). Scrolling down I found plenty of data. To check I converted the json file to ROOT and plotted a parameter, with the mean and # entries matching between the converted and original ROOT files. So, all seems well, closing.

Chris R, nothing at all, I'm putting the SciFiHit data in the existing SciFiHit branch in MC, as the data structure specifies. The move was made to preserve the data structure in fact ;)

Chris R, nothing at all, I'm putting the SciFiHit data in the existing SciFiHit branch in MC, as the data structure specifies. The move was made to preserve the data structure in fact ;)

Hurrah!

Looks like we tripped up because MCDigitzation is creating a null entry for every real entry. This should be fixed.

Bug found. Turns out the data coming from the MC supplied recon events, which were empty apart from the particle number. I was appending new recon events for the SciFiDigits, rather than appending the SciFi data to the existing recon events. Now fixed.

Maybe you don't have the recent enough version - I only added the root stuff in the last few weeks. Should be 0.14.0...