How to group dataframe rows into list in polars group_by

Question

import polars as pl

df = pl.DataFrame({
    "Letter": ["A", "A", "B", "B", "B", "C", "C", "D", "D", "E"],
    "Value': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
})

I want to group Letter and collect their corresponding Value in a List.

Related Pandas question: How to group dataframe rows into list in pandas groupby

I know pandas code will not work here:

df.group_by("a")["b"].apply(list)

TypeError: 'GroupBy' object is not subscriptable

Output will be:

┌────────┬───────────┐
│ Letter ┆ Value     │
│ ---    ┆ ---       │
│ str    ┆ list[i64] │
╞════════╪═══════════╡
│ A      ┆ [1, 2]    │
│ B      ┆ [3, 4, 5] │
│ C      ┆ [6, 7]    │
│ D      ┆ [8, 9]    │
│ E      ┆ [10]      │
└────────┴───────────┘

Answer 1

You could do this.

import polars as pl

df = pl.DataFrame(
    {
        'Letter': ['A', 'A', 'B', 'B', 'B', 'C', 'C', 'D','D','E'],
        'Value': [1, 2, 3, 4, 5, 6, 7, 8, 9,10]
    }
)
g = df.group_by('Letter', maintain_order=True).agg(pl.col('Value'))
print(g)

This will print

┌────────┬───────────┐
│ Letter ┆ Value     │
│ ---    ┆ ---       │
│ str    ┆ list[i64] │
╞════════╪═══════════╡
│ A      ┆ [1, 2]    │
│ B      ┆ [3, 4, 5] │
│ C      ┆ [6, 7]    │
│ D      ┆ [8, 9]    │
│ E      ┆ [10]      │
└────────┴───────────┘

maintain_order=True is required if you want to order of the groups to be consistent with the input data.