optional unique argument per field

[ThierryO]

unique fields cannot contain duplicated values

[jranke]

I like the concept of the git2rdata package, but have not tested it. In previous times, I have stored SQL databases as SQL dumps in version control, which has a similar effect but without any functionality to keep diffs as short as possible for data containing equivalent information.

Coming to this issue, why do you think it should be the job of git2rdata to enforce constraints such as unique fields? Couldn't such functionality be left to tools for relational data like dm?