Flattening multiple related rows in Power Query

[jPinhao]

We are loading some JSON data, which holds some arrays with varying property-value pairs. Due to the JSON structure, once you start expanding the fields you end up with something along these lines:

id | key1 | key2 | key3 | key4
0  | a    | null | null | null 
0  | null | b    | null | null
0  | null | null | c    | null
1  | a    | null | null | null
1  | null | b    | null | null
1  | null | null | c    | null
1  | null | null | null | d

We'd like to flatten these rows by ID to end up with something like:

id | key1 | key2 | key3 | key4
0  | a    | b    | c    | null 
1  | a    | b    | c    | d

The simplest way I found to solve this is by unpivotting the key columns, and then pivotting it back. However in some cases, we actually have quite a few more columns, and thousands of rows, in a quite sparse table. From previous trials I found that unpivot/pivot can be a big performance hit.

Is there a better way to solve this issue? Either in Power Query or with an R script?

EDIT: I beleive I found a way to do this with Table.Group:

Table.Group(sourceTable, {"id"}, {{"key1", each List.RemoveNulls([key1]){0},  type none}, {"key2", ...},  ...})

I'd like to do this across all columns (except ID), or from a List of column names - how would I go about making Table.Group work from a list of values dynamically?