I have an Explorer dataframe with a column that is a list type ({:list, {:s, 64}} ), and I’d like to mutate the dataframe to have a column that contains the sorted version of this list.
iex> df_with_list_col = DF.new(%{
list_col: [[1, 5, 3], [1, 1, 2, 0]],
regular_col: ["regular", "column"]
})
iex> DF.print(df_with_list_col)
"""
+-------------------------------------------+
| Explorer DataFrame: [rows: 2, columns: 2] |
+---------------------+---------------------+
| regular_col | list_col |
| <string> | <list[s64]> |
+=====================+=====================+
| regular | 1 |
| | 5 |
| | 3 |
+---------------------+---------------------+
| column | 1 |
| | 1 |
| | 2 |
| | 0 |
+---------------------+---------------------+
"""
I wanted to reach for a Explorer.Series list op , since there appears to be what I want in Polars, but it seems to be missing from the Explorer API. I was expecting to be able to do something like:
iex> DF.mutate(df_with_list_col, list_col_sorted: sort(list_col))
|> DF.print()
"""
+----------------------------------------------------------------------+
| Explorer DataFrame: [rows: 2, columns: 3] |
+---------------------+---------------------+--------------------------+
| regular_col | list_col | list_col_sorted |
| <string> | <list[s64]> | <list[s64]> |
+=====================+=====================+==========================+
| regular | 1 | 1 |
| | 5 | 3 |
| | 3 | 5 |
+---------------------+---------------------+--------------------------+
| column | 1 | 0 |
| | 1 | 1 |
| | 2 | 1 |
| | 0 | 2 |
+---------------------+---------------------+--------------------------+
"""
Is there perhaps a different recommended way of doing this? I suppose there is one hint here in this open issue, but I’m wondering if there is something much simpler and more obvious that I’m just missing.
