Samples with very few events per file
In https://gitlab.cern.ch/lhcb-datapkg/AnalysisProductions/-/issues/67 we're seeing a lot of failures due to no output data being found.
The cause seems to be because there are fewer than 100 events be LFN therefore a large fraction of files end up having no candidates of interest. This should be fixed in DaVinci (lhcb/DaVinci#78) but regardless I think it makes sense to have a merging step on these productions.
See also mc-requests#190 (closed)