Data Lineage Viz

Visualize SQL dependencies right inside VS Code. Ask @lineage in Copilot Chat to explore your graph with natural language — or browse interactively with search, trace, and schema overview.

Import from .dacpac files or connect directly to SQL Server, Azure SQL, Fabric Data Warehouse, or Synapse Dedicated SQL Pool.

Get Started

Run Data Lineage: Open Wizard (Ctrl+Shift+P)
Pick a .dacpac file — or Connect to Database via the MSSQL extension
Select schemas and click Visualize

No data? Click Try with demo data to explore the AdventureWorks sample.

AI-Powered Exploration

Use @lineage in GitHub Copilot Chat to query your loaded graph. The assistant uses dedicated lineage tools — never general knowledge.

@lineage what does HumanResources.Employee depend on?
@lineage trace 3 levels upstream from Sales.SalesOrderDetail
@lineage which objects are hubs with more than 10 connections?

Requires GitHub Copilot. Tools activate automatically when a graph is loaded.

Features

Search & trace — find objects with autocomplete, trace upstream/downstream dependencies, find shortest paths between nodes
Graph analysis — detect islands, hubs, orphans, circular dependencies, and longest chains
Schema overview — large graphs auto-summarize at schema level; double-click to drill in
SQL preview — click any node to view DDL with syntax highlighting; search across procedure and view bodies
Multiple sources — SSDT and SDK-style .dacpac, live database connections, external tables, virtual external refs (OPENROWSET, cross-DB, CETAS)
Projects & views — save connections, schema selections, and named filter states for one-click reopen
Export — generate Draw.io diagrams for documentation

How It Works

Extract — reads model.xml from a .dacpac, or imports metadata via DMV queries
Parse — combines XML/DMV dependencies with configurable regex patterns
Graph — builds a directed graph with automatic layout
Render — interactive visualization with pan, zoom, and filtering

Configuration

Search dataLineageViz in VS Code Settings (Ctrl+,).

Setting	Default	Description
`maxNodes`	`750`	Objects to load (up to 10,000). AI queries the full model; GUI is separately capped
`renderLimit`	`750`	Max nodes the GUI renders before showing a summary
`excludePatterns`	`[]`	Regex patterns to exclude objects by name
`layout.direction`	`"LR"`	Graph direction: left-to-right or top-to-bottom

Advanced users can override parse rules, DMV queries, and review profiling patterns. Scaffold templates via Command Palette: Create Parse Rules / Create DMV Queries.

Limitations

Object-level only — no column-level lineage
Static analysis — dynamic SQL (EXEC(@sql)) not detected
Fully-qualified names only — unqualified references are excluded

FAQ

Do I need a .dacpac file? No — connect directly to a database. If you prefer a .dacpac, extract one from Visual Studio, SSMS, Azure Data Studio, or the Fabric portal. See Microsoft's documentation.

Why are some dependencies missing? Dynamic SQL cannot be analyzed statically. Only compile-time dependencies are detected.

Contributing

Bug reports welcome. For custom features, fork and extend under the MIT license. See CONTRIBUTING.md.

MIT License · Christian Wagner · GitHub · Developed with Claude Code