Making datasets discoverable
Making a dataset discoverable allows users to find and view basic information about it without accessing its contents. Users can see the summary, description, creator, contributors, discussion, and license, but they cannot access any dataset files unless explicitly granted permission.
To help users determine whether a dataset is relevant to them, it is a best practice to:
Provide a detailed summary and description.
Enable file previews for selected files to showcase relevant data samples.
Important
If the dataset has any files explicitly marked as Preview files, those files will be visible to the users.
Setting a dataset to be discoverable
Only datasets owned by an organization (not by an individual user) and set to private can be made discoverable. When an organization owns a private dataset, it is not automatically shared with other organization members unless access is granted.
You can make datasets discoverable for the following users.
All members of an organization.
Specific members of an organization.
Users outside of the organization. In multi-tenant installations, anyone in the community can search for and access a discoverable dataset. In private installations, the dataset becomes visible to all other users within the installation.
To make a dataset discoverable:
On the Datasets details page, go to the Settings tab, and the Access and ownership section.
To make the dataset discoverable for all members of the organization, click the Grant access to all members link and select Discover.
To make the dataset discoverable for specific members, click the Grant access button and in the Grant access window, add the members and select the Discover option.
To make the dataset discoverable for anyone with access to the application, click the Share with community button (for multi-tenant installations) or the Share with everyone at your company button (for private installations).
When a dataset is discoverable, users who want full access must request permission from the dataset admins.
Making files available for preview
A file preview provides users with a sample of the dataset on the dataset overview page.
Enabling file preview helps users see a sample of the dataset before requesting full access. By default, this feature allows users to view the first five lines of a file, but they can still download the complete file unless additional restrictions are applied. To restrict access to sensitive data, use query-based previews to limit or mask certain information.
To enable a file preview:
Locate the file you want to make available for preview.
Click the Three-dot menu and select Edit metadata.
In the file settings, select Allow anyone to view this file when dataset is set to discoverable.
Click Save. The file is now labeled as preview and visible to the users with Discoverable access level.
Warning
Even though files marked as Preview cannot be accessed directly until permission is granted, any file visible on the overview page can still be downloaded by users who can see it.
Configuring secure and customized file previews
To provide a file preview, you can allow a random selection of 5 lines to be visible, but this setting permits users to download the entire file. If the file contains sensitive data or shouldn't be fully accessible, it's crucial to create an extract for the preview to restrict full access.
For files with sensitive information or large datasets, the default preview setting may not be suitable. Instead, employ a query-based extract to specify which data is displayed. For instance, if a file is tabular and holds a significant amount of data, you can run a query to display only a small portion, such as 5 rows, to limit the preview. Save these query results as a dataset using the Live View or Data Extract options, ensuring data visibility is appropriately controlled while offering a useful preview.
To create a custom preview file:
Locate the file you want to use as a custom preview file.
Click the Three-dot menu and select Query. A new query editor opens in the workspace.
Modify the query to limit the number of rows available for preview. For example, to display only five rows, use the following query.
SELECT * FROM retail_order_completeness LIMIT 5;For more details on using the SQL queries, see the Querying data page.
If you need to mask sensitive information, hide or anonymize specific fields in the file. For example, to replace sensitive values while keeping specific essential information, use the following.
SELECT date, 'XXX-XXX-XXXX' AS table, column, score FROM retail_order_completeness LIMIT 5;
This query hides the original table values, keeps the column and score fields, and limits the preview to five rows. For more details, see the Masking data for discoverable datasets page.
Next, click Run query to generate results. Click Download > Save to dataset or project. The Choose how data updates window opens.
Select a data saving method and click Save. Note: If you need to apply custom data types to table columns, use Data extract, as custom types cannot be applied to live data.
Notice
Name the result file appropriately to indicate that it is a preview extract, not the full dataset file.
Go to the saved file and enable preview, following the steps in the Making files available for preview section.