Dealing with missing values in DataFrames

[asinghvi17]

Currently, missing values in DataFrames (or DataFrames with column datatypes Union{Any, Missing}) cause errors in StatsMakie. What would be the best way to deal with this?

I think the best solution would be to use DataFrames.jl's integrated disallowmissing or dropmissing method, but that may cause unexpected behaviour (to the user).

[mkborregaard]

I don't think disallowmissing should be the first choice. missing values are part of Base Julia now, and they should be first-class citizens, treated mostly like NaNs.

[piever]

I think there are two possible ways forward:

1. Drop lines where one of the columns we use has a missing value (possibly with a warning)
2. Same as option 1. except when the users uses categorical axes (meaning, where under the hood we replace categories with numbers from 1 to n and add ticks to the axes based on categorical values), in which case missing should have its own category