SAP AI SDK for JavaScript - v2.17.0
    Preparing search index...

    Type Alias SamplingHeuristicConfig

    SamplingHeuristicConfig: {
        allowTimeoutFallback?: boolean;
        chunkSize?: number;
        chunkThreshold?: number;
        fuzzyDateDays?: number;
        fuzzyNumPct?: number;
        methodRatio?: number;
        poolSize?: number | null;
    } & Record<string, any>

    Representation of the 'SamplingHeuristicConfig' schema.

    Type Declaration

    • OptionalallowTimeoutFallback?: boolean

      When true, heuristic sampling falls back to random on timeout instead of raising an error. Set by the orchestrator when strategy=AUTO.

    • OptionalchunkSize?: number

      Max CASE WHEN expressions per subquery chunk (avoids HANA parse-tree depth limit). Default: 100.

    • OptionalchunkThreshold?: number

      Column count above which chunked SQL is used instead of a flat query. Default: 150.

    • OptionalfuzzyDateDays?: number

      Day window for fuzzy date matching. A date column scores 1 if abs(DAYS_BETWEEN(row_date, query_date)) <= dateFuzzyDays. Set to 0 for exact date match. Default: 30.

    • OptionalfuzzyNumPct?: number

      Numeric fuzzy threshold as a fraction of the absolute query value. E.g. 0.1 = ±10%. Default: 0.1.

    • OptionalmethodRatio?: number

      Fraction of numRows filled by scoring method (remainder is random fill). Must be between 0 and 1. Default: 0.67.

    • OptionalpoolSize?: number | null

      If set, scoring runs against a random subsample of this many rows instead of the full table. None = full table scan.