Polars groupby concat on multiple cols returning a list of unique values

Have a polars dataframe with the following data.

df = pl.DataFrame({"small": ["Apple Inc.", "Baidu Inc.", "Chevron Global", "Chevron Global", "Apple Inc."], "person": [10, 20, 30, 10, 10], "comp_big": ["Apple", "Baidu", "Chevron", "Chevron", "Apple"]})

Able to groupby on person, but this returns a dataframe of 2 lists.

(
    df
    .group_by("person")
    .agg(
        pl.col("small", "comp_big").unique()
    )
)

Returns

This is close, but want to merge the result between small and comp_big into a single list.

(
    df
    .group_by("person")
    .agg(
        pl.col("small") + "|" + pl.col("comp_big")
    )
)

This combines both into a single list, but I would want to split on the pipe and get the unique values.

How can this be achieved?

You can use pl.concat_list() and then .flatten()

(
    df
    .group_by("person")
    .agg(
        pl.concat_list("small", "comp_big")
          .flatten()
          .unique()
    )
)

shape: (3, 2)
┌────────┬───────────────────────────────────┐
│ person ┆ small                             │
│ ---    ┆ ---                               │
│ i64    ┆ list[str]                         │
╞════════╪═══════════════════════════════════╡
│ 20     ┆ ["Baidu", "Baidu Inc."]           │
│ 30     ┆ ["Chevron", "Chevron Global"]     │
│ 10     ┆ ["Chevron", "Apple Inc.", … "Che… │
└────────┴───────────────────────────────────┘