# Additional settings ## Test a contract on a sample {% hint style="warning" %} Currently, **this feature is only supported in Snowflake data sources**. {% endhint %} When testing a data contract, Soda allows you to **run contract validation on a sample of your dataset** instead of the full data. This feature helps you quickly and cost-efficiently verify that your contract runs correctly before executing full scans. Running a test contract on a sample enables you to: * Validate that your contract **syntax**, **checks**, and **filters** work as expected. * **Reduce data warehouse compute cost** while verifying new or updated contracts. * **Iterate faster** on contract definitions in development environments. Results from sampled runs reflect only a subset of your data and **may not represent its actual quality**. Use full verification once your contract logic is validated. ### Enable sampling for test contracts This feature can be enabled at the **data source level**, applying to **all datasets** that use that connection. {% hint style="warning" %} You need the "**Manage data sources**" global permission to add a new data source. Learn more about [Global and Dataset Roles](https://docs.soda.io/organization-and-admin-settings/global-and-dataset-roles). {% endhint %} To enable this feature: 1. Go to **Data sources**. 2. Click **Edit connection** for a data source.

3. Under the **Connection Details** section, toggle **Data Sampling**. 4. Specify your **sample size** on the **Limit** field.

5. Click **Connect**. *** ## Optimize computing with multiple warehouses {% hint style="warning" %} Currently available in **preview**. This feature is **only supported in Snowflake data sources**. {% endhint %} When connecting to Snowflake, you must provide a **warehouse** as part of the data source configuration. By default, this single warehouse is used for **all operations**, including discovery, metric monitoring, profiling, data contract executions, and the diagnostics warehouse. The **Configure warehouses per dataset** feature gives you greater control and flexibility by allowing you to **define specific warehouses for individual datasets**. This helps you **optimize cost, manage compute workloads, and allocate resources efficiently** across your data operations. {% hint style="info" %} This feature is available **only when using Soda Agent**. When using **Soda Core**, the warehouse can be specified directly in the **connection YAML** instead. {% endhint %} ### Enable the use of multiple warehouses {% hint style="warning" %} You need the **“Manage data sources”** global permission to enable or modify this feature. Learn more about [global-and-dataset-roles](https://docs.soda.io/organization-and-admin-settings/global-and-dataset-roles "mention"). {% endhint %} 1. Go to **Data sources** in Soda Cloud. 2. Click **Edit connection** for your **Snowflake** data source.

3. Toggle on **Configure Warehouses**. 4. Specify the **list of allowed warehouses** that can be used by this connection. 5. Choose a **default warehouse** to use for all datasets unless otherwise specified.

6. Click **Save** on the top right to save your configuration. ### Default warehouse behavior Once enabled: * The **warehouse** specified in the data source connection is used for **discovery**. * The **default warehouse** (defined under **Configure Warehouses**) is used for: * Metric monitoring * Profiling * Data contract executions * Diagnostics Warehouse operations * A different warehouse can be configured at the **dataset level**, overriding the default. ### Specify a warehouse at the dataset level {% hint style="warning" %} You need the **“Configure dataset”** permission to edit dataset-level configurations. Learn more about [global-and-dataset-roles](https://docs.soda.io/organization-and-admin-settings/global-and-dataset-roles "mention"). {% endhint %} 1. Go to a dataset in Soda Cloud. 2. Click **Edit dataset**.

3. Under the **Snowflake** section, select the **warehouse** to use for this dataset.