Cross-System DQ - Reconciliation Rule – Help Center

Overview

The Reconciliation rule is a Cross-System Data Quality (DQ) rule used to compare data between two tables, either at the record level or the aggregate level, within a single source or across different sources.

This rule allows users to:

Match records between tables using one or more key fields
Compare values between matched records
Validate record counts between datasets
Compare aggregated values

Creating a Reconciliation Rule

To create a Reconciliation rule:

Navigate to a table
Click Add Rule
Select Cross System DQ
Choose Reconciliation

Rule Configuration

After selecting the Reconciliation rule, configure the rule using the sections described below.

Reconciliation Type

The Reconciliation Type defines how records from the source and reference datasets are aligned for comparison.

Reconciliation types are grouped into:

Record-level checks
Aggregate checks

Comparison Type

The Comparison Type determines how records from the source and reference tables are matched during reconciliation.

Full
Includes all records from both source and reference tables, whether matched or unmatched.
Inner
Includes only records that exist in both source and reference tables.
Left
Includes all records from the source table and matching records from the reference table.
Right
Includes all records from the reference table and matching records from the source table.

The selected comparison type controls which records appear in the reconciliation results and how unmatched records are handled.

Record Level Check

The Record Level Check compares records between the source and reference tables by first matching records using defined key fields and then evaluating field-level values—including numeric, string, and date/time fields—for each matched record.

Available configuration sections:

Data Source & Table Selection
Choose the Source and Reference datasets for this comparison.
Key Fields Mapping
Select one or more fields that uniquely identify records for comparison..
Comparision Fields
Select fields whose values will be compared.
Context Fields
Select additional fields to include in context only. These fields are not used for comparison.

Aggregate Check

Compares aggregated values between the source and reference datasets.

Available configuration sections:

Data Source & Table Selection
Choose the Source and Reference datasets for this comparison.
Key Fields Mapping
Select one or more fields that uniquely identify records for comparison.
Comparison Fields
Select fields whose values will be compared.
Context Fields
Select additional fields to include in context only. These fields are not used for comparison.

Data Source and Table Selection

For all reconciliation types, users must define:

A Source connection and table
A Reference connection and table

Filters can be applied independently to both tables to control which records participate in the reconciliation.

Key Fields

Key fields define how records are matched between the source and reference tables.

By default, records are matched using the same field or set of fields selected on both the source and reference tables.

When record uniqueness depends on more than one field, multiple fields can be selected to form a composite key.

Asymmetric Keys allow matching records between the source and reference tables using different field combinations, enabling comparison even when no common key exists

Comparision Fields

Defines the source and reference fields to be compared after records are matched using key fields. Field selection is limited to compatible data types based on the chosen comparison type.

Comparision Type

Defines the comparision logic applied to the selected fields:

Equality – values must match exactly
Equality Null Match – treats values as a match when both values are equal or when either the source or reference value is null. Null values are not considered mismatches.
Difference – compares values within a specified tolerance
Absolute Value Difference – compares values using absolute tolerance
Field Selection
- Select the corresponding Source and Reference fields to be compared.
- For Difference and Absolute Value Difference, only numeric fields are available for selection.

Aggregation (Applicable for Aggregate Check only)

Defines the aggregation function applied before comparison:

SUM – calculates the total

MIN – selects the minimum value

MAX – selects the maximum value

AVG – calculates the average

The selected aggregation is applied to both source and reference fields before comparison.

Variance tolerance defines the acceptable difference between source and reference values when using Difference or Absolute Difference comparison.
- For numbers: can be set as a value or percentage
- For date/time: can be set in days or seconds

NOTE: Variance tolerance is not applicable for strings and If the difference is within the defined tolerance, the values are treated as a match

Context Fields

Additional fields included in results for reference.

These fields:

Do not participate in matching
Do not affect comparison logic
Help provide context when analyzing mismatches

Output Configuration

System Fields

These fields do not affect rule execution, scoring, or calculations. They act as derived, informational outputs that summarize comparison results, improving clarity for reporting, investigation, and governance workflows.

Filter Results

The Filter section controls how reconciliation results are filtered before the results view is loaded.

Users can define one or more conditions to filter the results based on selected columns and operators.

Note: Custom expression filters are not supported for the Reconciliation rule

Sort Order

The Sort Order section controls how reconciliation results are sorted when the results view is loaded.

Users can define one or more columns to sort the results and specify the sort direction for each column.

Sort Column
Select one or more columns to sort the results. Multiple columns can be added for sorting in sequence.

Custom Expression
Allows sorting results using a custom expression instead of a selected column.

Note: If no sort criteria are defined, the results are displayed using the default system order.

What You Can Achieve with the Reconciliation Rule

Using the Reconciliation rule, users can:

Compare records and values across systems
Validate data completeness and accuracy
Identify mismatches at both record and aggregate levels
Reconcile transactional or metric-based datasets using flexible matching logic

Cross-System DQ - Reconciliation Rule

In This Article