add `list.filter()`

[mcrumiller]

Problem description

filter would be a useful addition to the list namespace, which could just be syntactic sugar for list.eval(pl.element().filter(pl.element()...))

s = pl.Series([
    [1, 2, 3, 4, 5],
    [1, 3, 7, 8],
    [6, 1, 4, 5],    
])

# current implementation
s.list.eval(pl.element().filter(pl.element() < 5))

# proposed syntax
s.list.filter(pl.element() < 5)

shape: (3,)
Series: '' [list[i64]]
[
        [1, 2, 3, 4]
        [1, 3]
        [1, 4]
]

[parkma99]

9425

[orlp]

This looks good to me but I do have a concern. Currently pl.element is defined as

Alias for an element being evaluated in an eval expression.

We will need to change this, and create some sort of 'list expression' concept, where list.eval and list.filter accept 'list expressions'.

[DeflateAwning]

Would love this syntactic sugar. I derived the pl.col(list_col_name).list.eval(pl.element().filter(~ pl.element().is_in([0, 1]))) from first principles, but would way rather have just used the applicable .list.filter(...), if it existed.

[DeflateAwning]

I've once again been brought here. The double use of pl.element() is a little cringe.

[mkleinbort-ic]

Yes, would be good

[ghaffarialireza]

still waiting for it