In the current process, the list of teams in teamname_mapping.csv only grows over time. At some point, some rosters simply are not relevant anymore. Even if their name would return at some point, we can just treat that as a new team. A similar case holds for old rosters. It can be valuable to keep them for a while, for historic purposes (e.g. when we would want to look back at roster scores over time and we want to unify across different rosters with the same core); but not permanently.
Let's think about what would make sense as a solution:
- Remove every teamname that is no longer present in one of three rankings?
- Start with a clean "old_rosters" column every run?
- Add a counter for how long particular teamnames and rosters haven't changed, and trim them if the number is above a threshold? (The first two solutions would be special cases of this with a threshold of 0)
In the current process, the list of teams in
teamname_mapping.csvonly grows over time. At some point, some rosters simply are not relevant anymore. Even if their name would return at some point, we can just treat that as a new team. A similar case holds for old rosters. It can be valuable to keep them for a while, for historic purposes (e.g. when we would want to look back at roster scores over time and we want to unify across different rosters with the same core); but not permanently.Let's think about what would make sense as a solution: