CLB: Decisions Rematch in dataset options breaks decisions

[yroskov]

@mdoering, you might be interested to know:

Geoff and I taken an experiment in attempt to understand why the same Systema Dipterorum (SD) version had so many broken decisions after the re-import. (see the issue CatalogueOfLife/testing#127 (comment))

Experiment 1. We have re-imported SD with turned off setting Rematch Decisions:

Result: there were no broken decisions(!):

Experiment 2. We have re-imported the same SD with turned on setting Rematch Decisions:

Result: there are over 5K broken decisions:
https://www.checklistbank.org/catalogue/3/dataset/1101/tasks