Search result explainability

When you search for data assets by column names, sensitive data flags (such as an aspect indicating PII), or business glossary terms, search results traditionally return only the parent table or dataset. Without knowing why a table matched your query, you must open each search result, navigate through its schema or metadata tabs, and inspect dozens or hundreds of columns to locate the match.

Search result explainability in Knowledge Catalog (formerly Dataplex Universal Catalog) addresses this problem by displaying matching metadata directly on search result cards in the Google Cloud console. When a search query matches an asset's schema columns, entry-level aspects, column-level aspects, or linked business glossary terms, Knowledge Catalog displays explainability tags between the asset title and description. With these tags, you don't have to open every entry to find relevant fields. Instead, you can evaluate why each resource matched your query and identify relevant columns at a glance.

This document describes how search result explainability works, its benefits, the types of explainability tags, and how to view explainability in the Google Cloud console.

Use cases and benefits

Search result explainability provides context and efficiency when discovering data across your organization. Explainability can help you do the following:

  • Assess relevance quickly: You can determine whether a returned data asset meets your requirements without opening the entry details page to inspect its schema or attached metadata.
  • Identify column-level matches: When you search for sensitive data flags (such as Personally Identifiable Information, or PII), business terms, or specific column names, explainability tags show the exact schema columns associated with those elements.
  • Reduce search friction: Displaying matching context directly on search result cards reduces the need to navigate back and forth between search results and individual asset pages.
  • Refine search queries: Feedback from matched tags helps you determine whether your query was too broad, letting you refine your search with specific qualifiers or metadata filters.

How search result explainability works

Search result explainability is handled in the Google Cloud console rather than by the search API.

When you submit a search query in the Google Cloud console, the following process occurs:

  1. The search API evaluates your query and returns matching entries.
  2. The Google Cloud console fetches the metastore representation of each returned entry.
  3. The Google Cloud console performs keyword matching between your search terms and the entry's metadata, such as column schemas, attached aspects, and business glossary term links.
  4. When metadata elements match your query terms, the Google Cloud console renders explainability tags on the search result card.

Because explainability tags are generated through keyword matching against entry metadata, they only reflect direct matches with your search terms. Resources or columns returned by the search API based on semantic similarity are not highlighted in explainability tags. For example, if you search for given_name, the search API might return a table containing a column named first_name based on semantic meaning. However, the first_name column is not indicated in the explainability tags because keyword matching does not recognize the two terms as identical.

The following table describes the categories of explainability tags:

Tag category Console label Description Example
Columns in schema Columns: Matches one or more column names in the schema of the resource. Columns: customer_id
Aspects on entry Aspects: Matches aspects attached to the resource at the entry level. Aspects: Data Governance • confidentiality: Restricted
Columns with aspects Columns with aspects: Matches aspects attached to specific columns within the schema. Columns with aspects: phone_number / Data Governance • contains_pii: true
Terms on entry Terms: Matches business glossary terms linked to the resource at the entry level. Terms: Customer Account
Columns with terms Columns with terms: Matches business glossary terms linked to specific columns within the schema. Columns with terms: user_email / Personally Identifiable Information

Tag display and limits

Search result explainability tags include the following interface elements:

  • Help tooltip: An info icon appears alongside the explainability tags. Hovering over the icon displays a tooltip explaining that the list shows which parts of the resource matched your search.
  • Tag limits and overflow: Each category displays up to five tags. If more than five items match within a single category, an overflow tag indicates the remaining count (for example, ...and 3 more).
  • Tag details on hover: Hovering over an aspect or glossary term tag displays a tooltip with the complete aspect path, type name, or term details.

View search explainability in the console

To search for data assets and view explainability tags, follow these steps:

  1. In the Google Cloud console, go to the Knowledge Catalog Search page.

    Go to Search

  2. In the search field, enter your query. For example:

    • To find assets containing a specific column, enter the column name. For example, customer_id or column:customer_id.
    • To find assets tagged with specific aspect metadata, enter the aspect value or predicate. For example, PII or aspect:classification.tier=PII.
    • To find assets linked to a business glossary term, enter the term name.

    Press Enter to run your search.

  3. In the search results list, observe the explainability tags rendered on each result card, located between the asset name and description.

  4. Optional: Hover over an individual tag to view additional metadata, such as the full aspect path or term display name.

  5. To inspect the complete schema and metadata, click the asset name to open the entry details page.

    In the Schema tab, matched columns and associated aspects or terms appear highlighted.

What's next