Merge MAGIC data subrun-wise is not needed?

[YoshikiOhtani]

In the last weekly meeting Julian gave me a comment that M1 and M2 files may contain different sets of events even if the subrun ID is identical. Then I realized that it could affect the stereo reconstruction of MAGIC-only data, since I usually merge MAGIC DL1 files subrun-wise beforehand, but it may lose some events wrongly recognized as mono events. The best option would be to merge the files run-wise and use them for the stereo reconstruction (which takes more time per job though), and in that case the option --subrun-wise in merge_hdf_files.py is not needed anymore.

[YoshikiOhtani]

In order to avoid a confusion in the analysis, I would suggest to remove this option but I'm not sure who actually use it for the analysis. Please let me know what you think.

[aleberti]

Hi Yoshiki, yes, better to remove it at this point. MARS indeed in superstar looks at the subrun before and after the current subrun to look for matching events. If merging run-wise gets all the subruns putting them in a single df and then does the matching, then it is the same behavior as MARS and should be the only one allowed.

[YoshikiOhtani]

OK thanks Alessio, then I will open a pull request to remove the option.