I would like to achieve the following via Polars expressions, as opposed to mapping the elements row-by-row, but I have not been able to figure out a way.

import polars

def foo():
    return 1 + 1

def bar():
    return 1 + 1

def baz():
    return 2 + 2

exprs = 1_000_000 * [
    "foo() == bar()",
    "bar() == baz()",
    "foo() == baz()",
    "foo() > bar()",
    "bar() < baz()",
    "foo() != baz()",
]

df = polars.DataFrame({"exprs": exprs})

df.with_columns(
    polars.col("exprs").map_elements(
        lambda x: eval(x), return_dtype=polars.datatypes.Boolean
    )
)

shape: (6_000_000, 1)
┌───────┐
│ exprs │
│ ---   │
│ bool  │
╞═══════╡
│ true  │
│ false │
│ false │
│ false │
│ true  │
│ …     │
│ false │
│ false │
│ false │
│ true  │
│ true  │
└───────┘

The issue being that foo() and bar() are non-trivial and so map_elements() takes some time to complete since the DataFrame is large.