Back to Datahub

README

metadata-ingestion/docs/sources/thoughtspot/README.md

1.7.04.2 KB
Original Source

Overview

ThoughtSpot is a search- and AI-driven analytics platform used to author Liveboards (interactive dashboards), Answers (ad-hoc queries), and Worksheets (semantic models) over an organization's data warehouse.

The DataHub integration for ThoughtSpot ingests Liveboards as dashboards, Answers and Visualizations as charts, and Worksheets and Tables as datasets. When the ingestion principal can enumerate Orgs, those entities sit under Workspace containers; otherwise the catalog is flat.

Lineage is captured at both the table and column level: from Visualizations and Answers back to their Worksheets, and from federated Worksheets out to the underlying warehouse (Databricks, Snowflake, BigQuery, and similar). The connector also extracts ownership, ThoughtSpot tags, optional view counts, and uses stateful ingestion to soft-delete entities that disappear between runs.

Concept Mapping

ThoughtSpot ConceptDataHub ConceptNotes
Workspace (Org)Container (Workspace subtype)Parent of Liveboards, Answers, and Worksheets. Requires system-org admin to enumerate via /orgs/search.
LiveboardDashboard (Liveboard subtype)Formerly called "Pinboard"; the REST API still uses pinboard in some legacy field names.
AnswerChart (Answer subtype)Standalone saved searches.
VisualizationChart (Visualization subtype)Charts/tiles embedded inside a Liveboard.
WorksheetDataset (Worksheet subtype)TS's semantic layer — joined logical model over connection tables.
Logical View / SQL ViewDataset (View subtype)TS-side SQL definitions on top of a connection's schema.
Logical TableDataset (Table subtype)Direct one-to-one mapping of a physical source table.
TagGlobalTagResolved to urn:li:tag:<name> with URL-encoded names.
Owner / AuthorCorpUserDATAOWNER ownership; UUIDs in the author.name field are skipped (not real usernames).
Column source referenceUpstreamLineage + FineGrainedLineageWorksheet→Table column-level lineage from TS's pre-resolved columns[*].sources field on metadata/search.
Federated connectionUpstreamLineage (cross-platform)Worksheet → external databricks / snowflake / bigquery / etc. URN with table- and column-level edges; configured per-connection via external_connections.
View countDashboardUsageStatistics / ChartUsageStatisticsCumulative lifetime counter from metadata_search include_stats. Emitted by default; set include_usage_stats: false to skip.