Function duplicate_rows 

pub fn duplicate_rows(
    lf: LazyFrame,
    keys: &[PlSmallStr],
    polars_streaming: bool,
) -> Result<DataFrame>

The rows the duplicate check counted: every row equal to another in every one of keys, copies together, most copies first. One pass, grouping as the check did.

Copies are equal in every key, so each group’s key is its rows: it is repeated as many times as it occurs rather than looked up in a second read.