Question about saving data: We have a few different python environments, which are running different versions of pandas. As one might expect, pandas tables saved in a metaflow run in one environment aren't accessible in another, trying causes an unpickling exception. Does anyone else have this problem, and if so, how do you deal with it?

Not exactly this problem but maybe u can seraliaz the DF in an intermediate format like parquet ?

That's the best solution we've been able to come up with so far. This works for serializable data objects (pandas tables), but is an incomplete solution in the general case.

a solution would be to use a common env (conda). You can do that with the conda decorator but it’s currently hard to reproduce that env. I am going to push out a prototype version of helping out with this at some point in the near future which may help but otherwise yes, saving it in a stable format is a good way to go