# Sum Anomaly Reporting

Source: https://userguide.qualytics.io/data-quality-checks/sum/deep-dive/anomaly-reporting/

How the Sum check reports violations: the Shape Anomaly it produces, what the two numbers mean, and why the anomaly carries no source records.

## Anomaly Messages

Sum reports a **Shape Anomaly**, never a Record Anomaly. The rule evaluates the column as a whole, so there is no violating row to attach a message to.

> **Example: Shape Anomaly**
>
> ```text
> <field> sums to <observed>, which is not the expected <expected>
> ```

Reading the message:

- **`<observed>`** is the total the check computed across the filtered rows.
- **`<expected>`** is the expected total configured on the check.

A whole total is written without decimals. Anything else is rounded to three decimal places with trailing zeros dropped, so `95.500` reads as `95.5`. A value too small to survive that rounding, such as `0.0004`, is written in full instead. Thousands separators are applied only to whole totals, so `12,480` keeps them and `12480.5` does not.

Comparing the two tells you the size and the direction of the gap, which is usually the fastest way to recognize what went wrong: a total that is short by exactly one row's worth points at a dropped record, while a total that is roughly double points at a load that ran twice.

When a filter is set, the Shape Anomaly message names it at the end of its first sentence, as `(applied to records where <expression>)`.

Because the anomaly belongs to the column rather than to a set of rows, the Source Records view has nothing to list. Pair the check with a per-row rule when you also need to see the records.

**Custom Anomaly Description** does not apply to this rule. The option replaces the message on Record Anomalies, and Sum emits a Shape Anomaly only.

## See Also

- **Examples**

  Three production scenarios with sample data, the resulting anomaly message, and the SQL equivalent of what the check evaluates.

  [Examples](https://userguide.qualytics.io/data-quality-checks/sum/deep-dive/examples/)
- **Best Practices**

  Guidelines for control totals, scoping with a filter, and choosing between Sum and a per-row rule.

  [Best Practices](https://userguide.qualytics.io/data-quality-checks/sum/deep-dive/best-practices/)
- **Permissions**

  The team permission each action needs: view, create, edit, archive, restore, and delete.

  [Permissions](https://userguide.qualytics.io/data-quality-checks/sum/deep-dive/permissions/)
- **How It Works**

  The complete reference: definition, field scope, the expected total, exact equality, NULL handling, filter behavior, and why coverage does not apply.

  [How It Works](https://userguide.qualytics.io/data-quality-checks/sum/deep-dive/how-it-works/)

*Last updated: September 22, 2026.*
