...
- Discard field, row, etc
- Validation - Filter out data that doesn't meet a validation requirement. e.g. it could be that v2 supports something that v3 doesn't, in a way that's more nuanced than simply ignoring the data. Or it could be something like "v3 supports X length fields, v2 has Y length fields".
- Normalization - As in the concept of data https://en.wikipedia.org/wiki/Database_normalization
ID Replacement
- Replace an ID with another. Usually from some structural difference in schemas.
...