Changing values and type of JuliaDB column

General Usage
1

As a newcomer to Julia and JuliaDB, I have a simple question. Given a JuliaDB table, how would I map a function to a column that changes the column values and type? For example:

import JuliaDB
import Dates

function col_to_dt(val)
        dt = Dates.Date(replace(val, "|" => "-"))
end 

t = JuliaDB.table(["2015|03|01", "2016|02|05"]; names = [:dt])
JuliaDB.select(t, :dt => x -> col_to_dt(x))

The above code does not change t.dt. It remains a string. Also, in this example, how would you use JuliaDBMeta.@apply to conduct a series of steps to alter a column’s values and type?

Added Later:
I see that the following creates a new table and saves it under the same name:
t = JuliaDB.transform(t, :dt => JuliaDB.select(t, :dt => col_to_dt))

Would this be the recommended approach?

Still, how could this be arranged within a JuliaDBMeta.@apply call? The following does not work:

import JuliaDBMeta
import JuliaDB
import Dates

t = JuliaDB.table(["2015|03|01", "2016|02|05"], [2,4]; names = [:dt, :num])
t = JuliaDBMeta.@apply t begin
        JuliaDBMeta.@transform(dt = JuliaDBMeta.@select(t, :dt => col_to_dt))
      end

It gives the error:
ERROR: LoadError: LoadError: column 2015|03|01 not found.