Search Not Returning Expected Results

The DataHub search functionality is not returning expected datasets, columns, or entities, or search results seem outdated. This commonly occurs due to Elasticsearch index lag, incorrect search queries, or filtering that's too restrictive.

You Might Be Asking:

• Why can't I find my dataset in search?
• How do I search for column names?
• Why are search results outdated?

1. Verify the entity was successfully ingested:

# Check if dataset exists via API
curl -X POST http://localhost:8080/api/graphql \
  -H 'Content-Type: application/json' \
  -d '{"query": "{ dataset(urn: \"YOUR_URN\") { name platform } }"}'

2. (NOT GENERALLY RECOMMENDED) Force Elasticsearch index rebuild if needed:

# Trigger re-index via system endpoint
curl -X POST http://localhost:8080/api/v2/system/reindex \
  -H "Authorization: Bearer YOUR_TOKEN"

# Or use the CLI
datahub system reindex

3. Use advanced search syntax:

# Search in specific fields
name:my_table
description:customer
tags:PII

# Use wildcards
*customer*
sales_*

# Column-level search
fieldPaths:email_address

4. NOTE: Using Quotes For Exact Matches:

In DataHub search, enclosing terms in double quotes (e.g., "foo bar") enforces exact matching, preventing further tokenization—this is the documented approach for exact phrase searches. However, there are known inconsistencies: for terms containing underscores or dashes (e.g., "datahub_schema"), quoted searches may not behave as expected and can return non-exact matches due to how these characters are tokenized. For best results, use quotes for exact matches, but be aware of these edge cases. If you need to search for names with spaces, always wrap the phrase in quotes (e.g., name:"foo bar" in the API).

5. Check and adjust search filters:

   • Clear all filters in the UI
   • Verify platform filter isn't excluding your data source
   • Check domain and tag filters
   • Ensure environment filter matches (PROD vs DEV)

6. For column-level search, ensure schema metadata is indexed:

source:
  config:
    # Ensure schema ingestion is enabled
    include_views: true
    include_tables: true
    # Schema metadata is required for column search

Search indexing can take 1-2 minutes after metadata ingestion. For large deployments, consider increasing Elasticsearch resources. Column-level search requires the full schema metadata to be ingested.

• Search Guide
• Elasticsearch Configuration