Implement methods like "get", "first", "last" from list namespace in Array namespace

[Marlin-Na]

Problem description

It seems that Array type lacks some essential access methods like get in list namespace.

It would also be helpful to implement a to_list method to convert an Array type to List type.

[ritchie46]

It would also be helpful to implement a to_list method to convert an Array type to List type.

There is, you can can cast.

[Marlin-Na]

A follow up request for this is that if it is possible to implement non-strict type cast from array to list.

Currently if I cast to List without explicitly specifying the inner dtype, it will be casted to List(Null). It can be quite surprising to users.

>>> df = pl.DataFrame({"foo": [[1,2], [1,None]]}, {"foo": pl.Array(2, pl.Int32())})
>>> df
shape: (2, 1)
┌───────────────┐
│ foo           │
│ ---           │
│ array[i32, 2] │
╞═══════════════╡
│ [1, 2]        │
│ [1, null]     │
└───────────────┘
>>> df.select(pl.col.foo.cast(pl.List))
shape: (2, 1)
┌──────────────┐
│ foo          │
│ ---          │
│ list[null]   │
╞══════════════╡
│ [null, null] │
│ [null, null] │
└──────────────┘

[robinvd]

Im also missing the eval function from lists, or is there a already a way to do this?

example to add 2 to every array element

>> df = pl.DataFrame({"a": pl.Series("a", [[1,2], [3,4]], dtype=pl.Array(width=2, inner=pl.Int64))})
>> df.select(pl.col("a").cast(pl.List(inner=pl.Int64)).list.eval(pl.element() + 2).cast(pl.Array(2, inner=pl.Int64)))
shape: (2, 1)
┌───────────────┐
│ a             │
│ ---           │
│ array[i64, 2] │
╞═══════════════╡
│ [3, 4]        │
│ [5, 6]        │
└───────────────┘