Filter by multiple items in lists?

Question

Given a DataFrame with a list column and a list of items not in the data frame:

df = pl.DataFrame({
    "sets": [
        [1, 2, 3],
        [4],
        [9, 10],
        [2, 12],
        [6, 6, 1],
        [2, 0, 1],
        [1, 1, 4],
        [2, 7, 2],
    ]
})

items = [1, 2]

Is there an efficient way to filter the table to only have rows where the list column value contains a) one of the items in the list and b) all of the items in the list?

Expected result for ALL:

shape: (2, 1)
┌───────────┐
│ sets      │
│ ---       │
│ list[i64] │
╞═══════════╡
│ [1, 2, 3] │
│ [2, 0, 1] │
└───────────┘

Expected result for ANY:

shape: (6, 1)
┌───────────┐
│ sets      │
│ ---       │
│ list[i64] │
╞═══════════╡
│ [1, 2, 3] │
│ [2, 12]   │
│ [6, 6, 1] │
│ [2, 0, 1] │
│ [1, 1, 4] │
│ [2, 7, 2] │
└───────────┘

Answer 1

You can pass multiple tests to the any/all horizontal methods.

ALL

items = 1, 2

df.filter(
   pl.all_horizontal(
      pl.lit(item).is_in(pl.col("sets")) for item in items
   )
)

shape: (2, 1)
┌───────────┐
│ sets      │
│ ---       │
│ list[i64] │
╞═══════════╡
│ [1, 2, 3] │
│ [2, 0, 1] │
└───────────┘

• ^{(note: .filter now accepts *args syntax, which can replace .all_horizontal)}

ANY

df.filter(
   pl.any_horizontal(
      pl.lit(item).is_in(pl.col("sets")) for item in items
   )
)

shape: (6, 1)
┌───────────┐
│ sets      │
│ ---       │
│ list[i64] │
╞═══════════╡
│ [1, 2, 3] │
│ [2, 12]   │
│ [6, 6, 1] │
│ [2, 0, 1] │
│ [1, 1, 4] │
│ [2, 7, 2] │
└───────────┘