Search
Select Star’s Search is designed to help you find any data objects (tables, columns, Tableau Workbooks, Mode Reports, Looker Explores, etc.) without even knowing exactly what they’re called.
Use search in Select Star to answer questions such as...
Watch this video to learn more about Select Star's universal search, or keep reading for more details:
Search Types
Keyword Search
Traditional token-based search that matches specific keywords or phrases entered by users, used in the Top Search.
Natural Language Search (Requires enabling AI Features & Settings)
Allows users to input queries in a conversational manner, as they would ask a question to another person in conversational English and show them on the Search Results Page.
Top Search
The search bar at the top of the page is always available to find or quickly navigate to any data asset.

Space works as a delimiter in the search, so if you search for order items
you'll get results with both "order" and "items", whereas if you search for order_items
you'll get results that are looking for "order_items" as one word.
Note that when you search for multiple words, the order of the search terms matters!


All search results are sorted by relevance and popularity, so you’ll see what’s being used the most on the top. You won't have to worry about someone's test tables or dashboards showing up at the top of the search.
Click on different tabs below the search bar to see results from a specific data source. For instance if you're looking for a Looker Explore, click the Looker tab and then select Looker Explore.

Search Results Page
After you've typed a query into the top search, hit Enter
/Return
on your keyboard to see a full search results page and browse more high level information at once.
The tabs available in the top search dropdown are separated into individual filters on the side of the page, under Asset Types.

From this view, you can also choose to display or hide other items, such as Tags and Owners, by adjusting the column filters.

Natural Language Search
NLS lets users search for data assets by asking questions in everyday language - no need to know specific keywords or table names. If a query isn’t classified as a natural language question, it automatically falls back to standard keyword-based search using ElasticSearch.
Alongside the search results, NLS displays an AI-generated response similar to a chatbot, summarizing relevant information from the matched data assets. Users can click a “Continue the Conversation” button to ask follow-up questions, enabling a more interactive and conversational search experience.

Example Queries
How is revenue calculated?
Where can I find shipping facility preferences?
Can you find any dashboards or tables that mention revenue?
What information do we have on retention?
Who can answer questions about the business segments dashboard?
Limitations & Roadmap
Currently, NLS supports English only.
The classification algorithm identifies NLS queries based on phrasing patterns. It currently works best with questions starting with words like: answer, can, find, give, how, list, share, what, where, who, write. We recommend starting your question with one of these words to ensure it is classified as a natural language search.
Future improvements will expand question understanding and language support.
Enabling Natural Language Search
Make sure the "Enable AI Features" setting is turned on in the AI Settings section of the Admin panel (AI Features & Settings).
Once enabled, NLS will be active on the main Search page. The top navigation search bar will continue to use keyword-based search.
Note: We still can disable this specific feature upon request for your organization. Contact your Select Star account manager or support team to request it.
Search Ranking
Select Star ranks search results on a number of factors, including:
match type (name, breadcrumb, description)
popularity
asset type
tag
Search results are ranked based on several factors such as match type, asset popularity, and tags. For NLS queries, results are primarily ranked by the AI model’s relevance scoring, but users can still apply sorting and filters as usual.
Tag Boost in Keyword Search
Customers put a lot of effort to tag and organize data, and we want that to work for you. That is why we have updated our search algorithm to provide a slight boost for data assets tagged with specific tags. Data Assets with a tags with the following icon will receive a positive boost in the search results:

Last updated
Was this helpful?