Working with Snowflake Data Clean Rooms in the web app¶
The web app of a Snowflake Data Clean Room provides an intuitive UI that allows business users to create and use clean rooms without worrying about code complexities.
This topic provides an introduction to tasks that you complete as you work with a clean room. It describes the actions of the provider who creates and shares a clean room along with the consumer who uses that clean room.
Create a clean room as a provider¶
The Clean Rooms page in the web app lets you, as a provider, manage the lifecycle of a clean room, including creating and sharing. To create and share a clean room:
In the left navigation, select Clean Rooms.
Select + Clean Room. The creation process contains 4 steps:
Use the Add Data step to name the clean room and select the tables that are being shared with the consumer. The name can be 80 characters maximum, case-insensitive a-z, 0-9, spaces, and underscores.
Use the Specify Join Policies step to enhance your data with 3rd-party data and select which columns the consumer can join on.
Use the Configure Analysis & Query step to define which analyses the consumer can execute in the clean room, and select the tables and columns that they can use in their analysis.
Use the Share Clean Room step to invite consumers to use the clean room to collaborate. You can also use the Enable Run Analysis & Query option to specify which collaborators can run analyses in the clean room.
Run an analysis as a provider¶
A provider can configure a clean room so they can run their own analyses in the clean room. When setting up a clean room to run provider analyses, the provider can specify whether the consumer can also run analyses in the clean room. If the provider specifies that consumers cannot run analyses, the consumers act solely as data providers in the clean room.
Providers use the Enable Run Analysis & Query option during the Share Clean Room stage of creating a clean room to specify who can run an analysis. To configure the clean room so the provider can run an analysis, the provider turns on this option for their own account, which is listed along with the consumers.
When the provider wants to run an analysis, they can do one of the following:
Select Clean Rooms from the left navigation, find the tile for the clean room on the Created tab, and select Run.
Select Analyses & Queries from the left navigation, and run an existing analysis or create a new one just like a consumer would.
For both methods of starting an analysis, the provider can select which collaborator has the data that they want to include in their analysis.
Important
If a consumer allows a provider to run an analysis on a template, the consumer, not the provider, is charged for the credits consumed by the provider’s analysis. After the consumer has allowed the provider to run analyses, the consumer must uninstall the clean room to stop incurring costs.
If a consumer wants to obtain an estimate of the number of credits consumed by the provider within a specific time period, they can
execute the following query, where -5 returns an estimate of the previous 5 days of compute consumption by the provider:
SELECT * FROM table(samooha_by_snowflake_local_db_dev.public.udtf(-5));
If you are a consumer who wants to monitor the analyses being run by the provider, see Monitor provider-run analyses.
Push provider-run analysis results to Snowflake account¶
A provider can push results from a provider-run analysis to their Snowflake account, where it can be used for activation.
Important
The consumer who is contributing their data to a clean room that is configured for a provider-run analysis must consent to the provider’s ability to push results to the provider’s Snowflake account. To provide this consent as a consumer, complete these steps:
Start installing the clean room that is configured for provider-run analyses.
Continue the installation process until you get to the Configure Analysis & Query step.
In the Activation Settings section, select Allow activation for clean room provider.
To push results of a provider-run analysis to your Snowflake account, complete these steps:
After running the provider-run analysis, select Activate in the Results section.
Select the name of your Snowflake account.
Select Push Data.
For information about viewing the pushed data when it becomes available in the Snowflake account, see View analysis results in the provider’s Snowflake account.
Limitations on provider-run analyses¶
When using the web app to run an analysis, the provider has the following limitations:
Not all templates are supported. Currently, the Audience Overlap & Segmentation and SQL Query templates are supported.
If the collaborators are in different clouds or regions:
Consumers must enable cross-cloud auto-fulfillment on their account.
Results for a provider-run analysis are returned based on the combined refresh frequency between both parties. Providers and consumers should coordinate so the refresh frequency of the provider application and the consumer listing are similar (for example, both have a frequency of 15 minutes). This ensures that results are returned promptly.
Install a clean room as a consumer¶
The Clean Rooms page in the web app lets you, as a consumer, install clean rooms that have been shared with you by a provider. To install a clean room:
In the left navigation, select Clean Rooms.
On the Invited tab, find the clean room and select Join.
Select the tables that you want to use to collaborate with the provider’s data.
Use the Join Policies field to define the relationship between your data and the provider’s data.
Select Next.
Run an analysis as a consumer¶
As a consumer, you can use either the Clean Rooms page or the Analyses & Queries page to run analyses in an installed clean room.
To use the Clean Rooms page to run a new analysis based on the types of analyses that the provider has made available in the clean room:
In the left navigation, select Clean Rooms.
On the Joined tab, find the clean room in the list and select Run.
Select the analysis type, and then select Proceed.
Add filters to the analysis. There are two reasons why filter values might not be available:
The column contains more than 20 distinct values.
The clean room was recently installed, and has not finished processing preview values for the column. You can re-run the analysis when these values become available.
Select Run.
Optional: Expand the Save Analysis & Query section to save the analysis for future use.
To use the Analyses & Queries page to run existing analyses or create and run a new analysis:
In the left navigation, select Analyses & Queries.
Do one of the following:
To run an existing analysis, use the filters to find the analysis and run it.
To create and run a new analysis based on the types of analyses that the provider has made available in the clean room, select + New Analysis & Query.
Select a warehouse for an analysis¶
You can select which warehouse you want to use to run an analysis. Increasing the size or changing the type of the warehouse can speed up the analysis.
Note
The type of template determines what type of warehouse you can select for the analysis. For example, some templates (like Audience Overlap) only allow regular warehouses while other templates (like the Audience Lookalike Modeling) only allow snowpark-optimized warehouses.
The option to select a different warehouse appears next to the Run button on a template. This option does not appear for all templates.
Be aware that increasing the size of a warehouse or using a Snowpark-optimized warehouse can increase the cost of running the analysis. For information about how credit consumption grows as you use a larger warehouse, see Warehouse size and Billing for Snowpark-optimized warehouses.
For a description of the warehouses that are available, see Warehouses.
If you are an administrator who wants to create additional warehouse options, see Add warehouse options.
View details about a clean room¶
You can obtain details about a clean room, including:
A Collaborator Summary tab that lists the templates in the clean room along with the tables and join columns of your collaborator.
A My Summary tab that lists your tables and join columns.
A Table Relations tab that lists the relationship between your tables and the tables of your collaborator (that is, how the tables are joined).
A Data Stats tab that provides the following metrics for your tables:
My table: Shows how many distinct identifiers belong to a certain group. Note that statistics are updated every 24 hours so there might be a delay between modifying the clean room and seeing the updated statistics. Also note that columns with more than 20 distinct values are not shown.
Overlap stats: A clean room with either the Audience Overlap & Segmentation or SQL Query templates will show overlap stats to the consumer. These statistics describe how many distinct identifiers (join columns) belong to a certain group based on the attribute columns enabled in the template. You can select up to 2 attribute columns to view statistics breakdowns for. The data is generated after the initial installation and is refreshed whenever a user logs into the clean room web app. Note that in the bar graph visualization only the first 5 rows of data are plotted based on the default sorting provided. Statistics that take longer than 10 minutes to run for a particular breakdown will not be available.
To access these clean room details, complete these steps:
In the left navigation, select Clean Rooms.
Click on the tile for the clean room.
Clean room details are also available in the Clean Rooms Details section when you are running an analysis.