Trouble creating new column in grouped object

Specific Domains Data
1 
using Query
using DataFrames

ex_df = DataFrame(id = [1,1,1,2,2,2,2], trip= ["a", "a", "b", "a", "a", "b", "b"], values=rand(7))

function consect(d)
    # 0 if consecutive values are the same, 1 otherwise
    Int.(view(d, 2:length(d)) .!== view(d, 1:length(d)-1))
end

ex_df |> 
    @groupby(_.id) |>
    @map({id=key(_), run=cumsum([1, consect(_.trip)...]), rows=_ }) |>
    @mapmany(i->i.rows, (i,j)->{j..., i.run}) |>
    DataFrame

7 rows × 4 columns
        id	trip	values	        run
	Int64	String	Float64	        Array…
1	1	a	0.430749	[1, 1, 2]
2	1	a	0.806467	[1, 1, 2]
3	1	b	0.744969	[1, 1, 2]
4	2	a	0.652944	[1, 1, 2, 2]
5	2	a	0.326956	[1, 1, 2, 2]
6	2	b	0.717417	[1, 1, 2, 2]
7	2	b	0.448909	[1, 1, 2, 2]

I want the run values to be unraveled so that there is a single integer per row instead of the full array. That is the values for run should be [1, 1, 2, 1, 1, 2, 2] instead of the current array of arrays.

How can I do this with Query? Thanks in advance!

2

I think

@mapmany(i->zip(i.rows, i.run), (i,j)->{j[1]..., run=j[2]})

might do the trick?

1 Like