Member Junction
    Preparing search index...

    Per-key statistics for CUSTOM (unmapped) source keys observed during a sync — aggregated IN MEMORY across every fetched record, REGARDLESS of whether the row was written or the content-hash fast path skipped it. Custom-overflow behaviour: a newly-appearing custom column must NOT affect the row-hash match (so unchanged rows stay skip-cheap), yet its presence + sizing statistics must still surface as a promotion candidate at sync end. This out-of-band aggregation is what makes both true at once — candidates and generous sizing stats exist even when every row was skipped.

    interface CustomKeyStat {
        Key: string;
        MaxLength: number;
        Occurrences: number;
        SampleValues: unknown[];
        TotalRecords: number;
    }
    Index

    Properties

    Key: string

    The unmapped source key as it appears in ExternalRecord.Fields.

    MaxLength: number

    Longest observed String(value).length — sizes the future column generously.

    Occurrences: number

    Records in which the key appeared with a non-null value.

    SampleValues: unknown[]

    Bounded sample of observed values, for type inference at promotion time.

    TotalRecords: number

    Total records scanned for this entity map this run.