Converting NamedTuple to DataFrame seems expensive?

[pdeffebach]

Use copycols = false in the DataFrame constructor.

If you can avoid a dependency on data frames that would be best.