weights not yet supported on array with null values

Error message

weights not yet supported on array with null values

What it means

nulls::rolling_sum, the null-handling rolling sum, panics with 'weights not yet supported on array with null values' when weights is Some. Only the null-free path implements weighted summation; the guard makes the combination an explicit failure rather than a silent wrong total.

Source

Thrown at crates/polars-compute/src/rolling/nulls/sum.rs:24

    arr: &PrimitiveArray<T>,
    window_size: usize,
    min_periods: usize,
    center: bool,
    weights: Option<&[f64]>,
    _params: Option<RollingFnParams>,
) -> ArrayRef
where
    T: NativeType
        + IsFloat
        + PartialOrd
        + Add<Output = T>
        + Sub<Output = T>
        + SubAssign
        + AddAssign
        + NumCast,
{
    if weights.is_some() {
        panic!("weights not yet supported on array with null values")
    }
    if center {
        rolling_apply_agg_window::<SumWindow<T, T>, _, _, _>(
            arr.values().as_slice(),
            arr.validity().as_ref().unwrap(),
            window_size,
            min_periods,
            det_offsets_center,
            None,
        )
    } else {
        rolling_apply_agg_window::<SumWindow<T, T>, _, _, _>(
            arr.values().as_slice(),
            arr.validity().as_ref().unwrap(),
            window_size,
            min_periods,
            det_offsets,
            None,

View on GitHub (pinned to 9b5d73fd00)

Solutions

Compute the weighted sum manually from supported ops: fill nulls, then rolling_sum(x * w) / rolling_sum(w) for a mean or rolling_sum(x * w) directly for a decayed total
Drop weights and use the null-aware unweighted rolling_sum
Fill nulls (e.g. fill_null(0)) so the no-nulls weighted kernel is used, documenting the bias
Request weighted null-aware rolling sum upstream

Example fix

# before
s.rolling_sum(window_size=7, weights=[0.5, 0.7, 0.8, 0.9, 1.0, 1.0, 1.0])  # nulls -> panic
# after
s.fill_null(0.0).rolling_sum(window_size=7, weights=[0.5, 0.7, 0.8, 0.9, 1.0, 1.0, 1.0])

Defensive patterns

Strategy: validation

Validate before calling

def safe_rolling_sum(s: pl.Series, window_size: int, weights=None, min_periods=1):
    if weights is not None and s.null_count() > 0:
        raise ValueError("weights + nulls unsupported for rolling_sum; fill first or drop weights")
    return s.rolling_sum(window_size, weights=weights, min_periods=min_periods)

Try / catch

try:
    out = s.rolling_sum(window_size=7, weights=w)
except pl.exceptions.PanicException:
    out = s.fill_null(0.0).rolling_sum(window_size=7, weights=w)

Prevention

null_count() check before any weighted rolling op - it covers sum, mean, var, min/max, quantile
Fill nulls explicitly (and record the choice) rather than letting a panic surface at query time
Wrap rolling_* in team helpers that enforce the weights/nulls rule once

When it happens

Trigger: s.rolling_sum(window_size=k, weights=[...]) or .rolling_sum(..., weights=...) on a Series containing at least one null value; rolling sums with decay weights over nullable logs/metrics.

Common situations: Weighted moving sums over event counts with missing intervals; code ported from pandas (which tolerates NaN + weights differently) that panics under polars once nulls occur.

Related errors

AI-assisted analysis of pola-rs/polars@9b5d73fd00 (2026-08-19). Data as JSON: /api/errors/12a270a4726df2aa. Report an issue: GitHub.