polars.Expr.fill_nan

Expr.fill_nan(value: int | float | Expr | None) → Self

Fill floating point NaN value with a fill value.

Examples

>>> df = pl.DataFrame(
...     {
...         "a": [1.0, None, float("nan")],
...         "b": [4.0, float("nan"), 6],
...     }
... )
>>> df.with_columns(pl.col("b").fill_nan(0))
shape: (3, 2)
┌──────┬─────┐
│ a    ┆ b   │
│ ---  ┆ --- │
│ f64  ┆ f64 │
╞══════╪═════╡
│ 1.0  ┆ 4.0 │
│ null ┆ 0.0 │
│ NaN  ┆ 6.0 │
└──────┴─────┘