Less Than Field Best Practices

Guidelines for getting reliable signal from Less Than Field checks while keeping the noise (and the maintenance) low.

Decide what an equal pair means

A single-day contract, or usage that lands exactly on the quota, are legitimate in most models. The Inclusive setting decides whether those rows pass, so make the call deliberately rather than discovering it from the first anomaly.

Compare against a column, not a snapshot of it

The point of Less Than Field is that the reference moves with the data. When you find yourself tempted to hardcode today's value of the other column, use Less Than instead and be explicit that the boundary is a constant.

Set a tolerance only for noise you can name

Two timestamps written by different systems, or two amounts that accumulate rounding, differ for reasons that have nothing to do with data quality. Those are worth a comparator. A margin added to silence an inconvenient anomaly hides the defect the check exists to catch.

Remember NULLs make the row pass

With no comparator configured, a row where either side is missing is never reported, so a broken upstream job that nulls a column also silences this check. With a comparator configured, only a row where both values are NULL passes. Pair the check with Not Null on both fields when their presence is part of the rule.

Pair with Not Null when the value is mandatory

The rule asserts that present values satisfy it; it does not require the field to be populated. When the field must also have a value, add a Not Null check on the same field.

Keep coverage at 100% unless a known backlog exists

At 100% coverage every failing row is reported as a Record Anomaly, which tells you exactly which values break the rule. Below 100% the check reports a single Shape Anomaly only when the failing fraction crosses the tolerance, and no per-row detail is produced. Lower coverage only while a known set of legacy rows is being cleaned up.

Scope with a filter instead of loosening coverage

When the rule only applies to part of the table (one product line, one tier, one period), express that with a filter clause rather than by lowering coverage. The filter removes the out-of-scope rows from evaluation entirely, and the expression is echoed in every anomaly message.

Route the anomalies to the right people

An out-of-bounds number usually comes from an application that failed to validate its input, or from a unit mismatch upstream. Set an Anomaly Assignee from the team that owns that producer, and tag the check so related checks are easy to find.