average
average(a, weights=None, axis=None, *, validate=True)Return the weighted average with NaN/inf propagation.
Parameters
a : array_like-
Input data. Supported numeric inputs are normalized to a contiguous kernel array when
validate=True. Computed reducers promote integer and bool inputs tofloat64; exact selection reducers keep integer and bool dtypes where the selected value can be returned exactly. Complex and object arrays are not supported. axis : None, 0, -1, or int = None-
Axis to reduce.
Nonereduces the whole array.0reduces strided reducing-axis slices into the remaining shape.-1andndim - 1reduce contiguous slices. Other axes raiseNotImplementedError. weights : None or array_like = None-
Weights for a weighted reduction. With
axis=None, weights must have the same shape asa. With an axis reduction, weights may either have the same shape asaor be 1-D with length equal to the reducing axis. validate : bool = True-
If
True, check dtype, dimensionality, contiguity, and axis validity before entering the Rust kernel. IfFalse, the caller must provide a contiguous supported kernel dtype (float32,float64, bool, or a NumPy integer dtype).validate=Falseskips dtype promotion: integer and bool arrays are reduced directly, while complex and object arrays remain unsupported.
Returns
out : float or ndarray-
Reduction result.
axis=Nonereturns a Pythonfloat. Axis reductions return an array with the reduced axis removed.
Notes
Plain reducers include every value, so NaN and inf propagate with IEEE / NumPy-like semantics. This is the fastest path for known-clean finite data.
Weighted averages return floating results, so integer and bool inputs are promoted. A zero sum of retained weights raises ZeroDivisionError. For nanaverage, NaN values in a are skipped (with ignore_inf=True, all non-finite values in a are skipped), and the weights attached to skipped values are skipped with them; a NaN weight on a retained value propagates to the result.