Choosing Validation scope and method
Before creating a Validation, you must decide what to validate (the Scope) and how Validata should perform the comparison (the Method).
Selecting the right combination ensures that the Validation configuration aligns with your objectives. Once selected, you cannot change the Scope and Method in your Validation configuration.
Validation scope and methods
Validata provides two distinct scopes:
Singleton Validation Pair (Quick Start):
You can quickly get started by validating a single Validation Pair—comparing one Source table against its corresponding Target table.
Flexibility: A Singleton Validation Pair focuses on a single table-to-table comparison, giving you full control over how that specific pair is validated without the complexity of managing a larger configuration. Once configured, you cannot add additional Validation Pairs to this validation instance.
Supported Methods: All built-in validation methods are supported at this scope—Vector Validation, Fast Record Validation, Full Record Validation, Key Validation, and Interval Validation—as well as Custom Validation, which is available only for single-table comparisons.
Best for: Quick testing, targeted investigations, and one-off validations where you want complete control over how a specific Source–Target table pair is compared.
Validation Set:
You can use the Validation Set scope to compare one or more Validation Pairs simultaneously, allowing you to validate tables across one or multiple schemas in a single run.
Flexibility: A Validation Set is designed for scale and adaptability. You can select tables manually or use wildcard patterns to define rules that automatically select matching tables. Unlike the Singleton scope, you can modify this configuration at any time by adding new Validation Pairs or deselecting existing Pairs.
Supported Methods: This scope supports the standard suite of built-in methods: Vector Validation, Fast Record Validation, Full Record Validation, Key Validation, and Interval Validation.
Best for: Larger or recurring validations where you want to compare multiple tables together, apply consistent validation rules across schemas, or automate broad data quality checks. A Validation Set is highly customizable but may take more time to refine than a single-table validation.
See the Concepts article on Validation Types to learn more about the different Validation methods.
Validation use cases and recommendations
Example use case | Recommendations & Considerations |
|---|---|
You want to compare a single Source table with its corresponding Target table. |
|
You want to compare data between two databases. |
|
You want to validate data in a data warehouse that stores data in APPEND mode (paired with a database or another warehouse). |
|
You want to validate data in a data warehouse that stores data in MERGE mode (paired with a database or another warehouse). |
|
You want to validate data in a data warehouse where some tables are populated in APPEND mode and other tables are populated in MERGE mode (paired with a database or another warehouse). |
|
You want to validate data that is transformed before being written to the Target. |
|
You want to match record counts between the Source and Target datasets. |
|
You want to run a Validation every few hours to validate data that was recently updated. |
|
You want to validate records between Source and Target table pairs where the comparison key is a text column that behaves differently between Source and Target due to differences in collation or character-set settings. |
|
Operational considerations
You can significantly enhance your validation strategy by combining multiple validation methods to achieve comprehensive coverage. A common approach is to use partial-dataset methods—like Key Validation or Interval Validation—to quickly detect incremental discrepancies, complemented by scheduling a full-dataset validation (e.g., Vector, Fast Record, or Full Record Validation) during your scheduled maintenance windows.