Add get function for AbstractDataFrame

[dgegen]

Analogous to the syntax of the df.get function in Pandas, I propose to add the following function in src/abstractdataframe/iteration.jl, which I syntactically consider a conventional and convenient way to deal with cases where the return of default values is desired.

"""
    get(df::AbstractDataFrame, col_ind, default=nothing)
Return the column with the index `col_ind`, or the given default value `default`
if the column of the given index is not present in the DataFrame `df`.
"""
function Base.get(df::AbstractDataFrame, col_ind::Union{Symbol,AbstractString},
                  default=nothing) 
    hasproperty(df, col_ind) ? df[:, col_ind] : default
end

Does anyone have additional suggestions, thoughts or opposing arguments?
Cheers!

[bkamins]

we could add:

get(df::DataFrameColumns, col_ind::ColumnIndex, default)

and

haskey(df::DataFrameColumns, col_ind::ColumnIndex)

so @dgegen could get the expected behavior, but with eachcol(df).

@nalimilan - what do you think about it?

[nalimilan]

IIRC we had this a long time ago, but we dropped it because it goes against the idea that data frames are collections of rows. It would have to be called getcols if we wanted to add it.

[bkamins]

@nalimilan - that is why I do not want to add it to DataFrame but to DataFrameColumns. DataFrameColumns is a collection of columns so it could support get and haskey.

Note that we now support both get and haskey for DataFrameRow for the same reason (as it is a collection of columns).

[nalimilan]

Woops, sorry. Why not then. :-)

[bkamins]

Added in #3282