Missing data

q.transform.impute provides sklearn-compatible fitted missing-data handling:

Transformer	Purpose
SimpleImputer	Univariate constant or summary-statistic imputation
KNNImputer	Nearest-neighbor imputation
IterativeImputer	Multivariate chained-equation imputation (MICE)
MissingIndicator	Binary missing-value indicators

Use these transformers inside q.transform.Pipeline so learned imputation state is fitted only on fit partitions:

pipeline = q.transform.Pipeline(
    [("impute", q.transform.impute.SimpleImputer(strategy="median"))]
)
imputed = pipeline.fit_transform(dataset)

The transformers are provided by scikit-learn and exposed through the QRT namespace for a consistent public API. Deterministic gap detection and data canonicalization belong in q.data.