Polars Rust equivalent to pl.lit() (repeated value in df)

Question

In python I can construct a dataframe with a repeated value like this:

import polars as pl
df = pl.DataFrame({"foo": [1,2]}).with_columns(bar=pl.lit("baz"))

Can this be done in rust? I'd expect something like:

use polars::prelude::*

let df = df!["foo" => [1,2]].unwrap().with_column(Literal::String("baz"));

but "the trait IntoSeries is not implemented for polars::prelude::LiteralValue". Looking at that trait I can't see any implementers which aren't already collections.

Is the supported way to do this in rust just to construct a vec/series of N repeated values and use that? Or is there some way to have polars handle the repetition itself?

Answer 1

The easiest solution would be to convert your DataFrame into LazyFrame(by calling .lazy() method). Then use the lit function to create a column with literal value.

use polars::prelude::*;

fn main() {
    let df = df!["foo" => [1,2]]
        .unwrap()
        .lazy()
        .with_column(lit("baz").alias("bar"))
        .collect()
        .unwrap();
    println!("{df}");
}

If you directly want to operate on DataFrame without converting into Lazyframe, You may need to do some additional work to get that done.

use polars::prelude::*;

fn main() {
    let mut df = df!["foo" => [1,2]].unwrap();
    df.with_column(Column::new_scalar(
        PlSmallStr::from("bar"),
        Scalar::from(PlSmallStr::from("baz")),
        df.height(),
    ))
    .unwrap();
    println!("{df}")
}