Validatar has a new release! From major updates to minor bug fixes, we’re constantly working to improve your Validatar experience so you can continue to build trust in your data. Learn about what’s new below.
Release Date
Validatar 2026.1 will be released to cloud environments on May 7, 2026.
What’s New!
Databricks Data Processing Engine
Validatar 2026.1 introduces full support for Databricks as a data processing engine, enabling organizations to run data quality tests and execute profile sets directly within their Databricks environment.
Test Engine
Connect Validatar to your Databricks SQL warehouse to execute data quality tests natively within Databricks without requiring any external data movement.
The engine uses the Databricks Statement Execution API with support for large result sets, allowing comparisons of arbitrarily large datasets.
Custom success criteria calculations use SQL expressions evaluated natively in Databricks.
Cloud instances notify Validatar when processing completes and results are available.
Profile Sets
Execute profile sets (table-level or column-level data profiling) against Databricks tables, with results stored in your Databricks account and viewable in Validatar alongside profiles from other engines.
Custom calculations in Databricks profile sets use SQL expressions evaluated natively in Databricks.
Data Source Template
A dedicated Databricks data source template type is now available, allowing teams to define reusable connection patterns for Databricks-backed data sources.
Data sources can also be created using a new Databricks Driver connection type, allowing access to Databricks without requiring a separate Databricks ODBC driver to be installed.
Repository Replication
Repository replication now supports Databricks as a destination engine, allowing replicated test data and configurations to be stored in a Databricks database alongside built-in and Snowflake destinations for reporting purposes.
AI Updates
New and Expanded Validatar AI Tools
Data Source Management: Get, create, and update data sources. Validatar AI can now inspect existing data source configurations and create new ones, enabling more complete end-to-end test setup workflows.
Create Tests: Create both standard tests and template-based tests
Execute Script: Execute a data source query/script on demand and retrieve results
List Tool Enhancements: The List tool now supports listing projects, reports, project folders, and folder objects in addition to standard/template tests.
Describe Tool Enhancements: Describe tool updates provide richer detail when Validatar AI inspects data sources, standard/template tests, and other objects.
Revert: Supports undoing prior actions taken by Validatar AI
MCP Server
Validatar now exposes its AI tools as an MCP (Model Context Protocol) server, allowing external agentic AI tools (such as Claude Code, Codex, Cursor, or other MCP-compatible clients) to call Validatar AI tools directly.
API tokens for MCP access can be scoped to specific tools. Token owners can configure each token to allow only a defined list of read-only tools, writable tools, or both, restricting what an external AI client can do with a given token.
MCP requests are logged with tool name and parameters, providing an audit trail of AI-driven operations.
New Report Datasets
User/Group Role Membership: A new report showing which users and groups are assigned to each role, making it easier to audit who has access to what.
Effective User Permissions: A new report showing the computed (effective) permissions for each user across objects, accounting for group membership and role inheritance.
Object Member Permissions: A new report showing which users and groups have permissions on a specific object type, useful for auditing access to individual projects or data sources.
New API Endpoints
Authoring Endpoints
Update Project
Get/Update Project Custom Fields
Update Data Source
Delete Custom Metadata Source
Delete Profile Set
Delete Custom Profile
Import/Export Data Source Template
Get Standard Test Current Version Detail
Get Standard Test Version Detail
Get Template Test Current Version Detail
Get Template Test Version Detail
Get/Update Test Custom Fields
Get/Update Job Custom Fields
Get/Update Label Custom Fields
Catalog Endpoints
Get/Update Data Source Custom Fields
Get/Update Schema Custom Fields
Get/Update Table Custom Fields
Get/Update Column Custom Fields
Reporting Endpoints
Create/Update Report
Other Highlighted Enhancements
Both Snowflake and Databricks data processing engines now support high levels of parallel processing, both for test execution and profile set execution. Prior to this release, Snowflake data processing engines had a limitation of 20 processes running in parallel.
The standard Data Source Templates (SQL Server, Snowflake, and PostgreSQL) have been updated with the following changes: 1) added leading/trailing delimiters for schema, table, and column names, and 2) updated the Column Metadata for a Table and List of Tables in a Schema macro scripts to account for leading/trailing delimiters. To overwrite these templates in your Validatar instance, follow these steps in Validatar:
Navigate to the Marketplace page and filter Included Content to Data Source Templates.
Click on each of the three templates to open the detail popup, click the Get button, then click the Import Directly button.
Check the Overwrite the existing data source template option, then click the Import Template button. NOTE: This will overwrite any changes you have made to the data source template in your Validatar instance.
Bug Fixes
Resolved an issue where API authentication tokens were not correctly invalidated when a token was changed or revoked.
Resolved an issue where project-level scoping with API tokens was not applied correctly to tokens linked to a global admin user. NOTE: This could cause a change in behavior where an API call will fail now when it previously succeeded due to incorrect project-level scoping being applied in prior versions.
Fixed an issue in the built-in and Snowflake processing engines where test comparisons involving NULL values produced incorrect results.
Fixed a macro issue with the built-in SQL Server, Snowflake, and PostgreSQL data source templates when using leading/trailing delimiters by introducing a new unquote helper method that can be used in macro templates.