[julia-users] Working with DataFrame columns types that are Nullable

[John Best]

I've got ODBC.jl set up to retrieve a couple of queries. This works, but it 
is returning a DataFrame with column eltypes of Nullable{Int64}, 
Nullable{Dec64}, etc. I'd like to convert the numeric element types to 
Float64 for use in my analysis (which was written based on reading .csv's 
of the data). I would try to present an example, but I can't seem to 
construct a basic DataFrame with NullableArray columns without getting:

    ERROR: MethodError: `upgrade_vector` has no method matching 
upgrade_vector(::NullableArrays.NullableArray{Int64,1})WARNING: Error 
showing method candidates, aborted

     in setindex! at 
/home/jkbest/.julia/v0.4/DataFrames/src/dataframe/dataframe.jl:368
     in DataFrame at 
/home/jkbest/.julia/v0.4/DataFrames/src/dataframe/dataframe.jl:104

This is also the error I get when I try to manually convert a column, i.e.

    df[:colA] = NullableArrays.NullableArray{Float64}(df[:colA])

This is the first time I've tried working with NullableArrays. Is there any 
way to convert a DataFrame of NullableArrays to a DataFrame of DataArrays? 
And how would I do that? At least I know that my existing code works for 
that.

Thanks,
John