​

Explore the Lineage View

Collate displays end-to-end lineage for the table and column levels, for Database, Dashboard, and Pipeline assets. Search for a data asset and expand the graph to see its lineage:

• Each node expands to show its upstream and downstream edges.
• Each edge carries details such as the SQL query, pipeline information, and column-level lineage.

​

Source, Target, and Edges

In the example below:

• The table on the left is the parent, or Source node.
• The table on the right is the Target node, identifiable by the arrow pointing into it.
• The arrow connecting them is the Edge.

Click an edge to see its details: the Source, Target, Description, and SQL Query. For a view, the SQL query shows how the target was generated from the source.

​

Lineage Config

Control how many upstream and downstream nodes appear in the graph. This only affects your current view: it resets to the platform default the next time you open a lineage graph. To configure the lineage nodes, follow the steps below:

1. Click the Setting icon.
2. Set the following fields:
   • Upstream Depth: Number of upstream hops shown by default when opening a lineage graph, to identify parent-level sources.
   • Downstream Depth: Number of downstream hops shown by default when opening a lineage graph, to identify child-level targets.
   • Nodes Per Layer: Maximum number of nodes shown per layer. There’s no fixed maximum here. If a layer has more nodes than this, the rest are paginated instead of all being rendered at once.
   

To change the default that applies for every session, see Lineage.

​

Filtering Lineage by Time

On an asset’s Lineage tab, use the calendar control in the toolbar to limit the graph to edges observed within a specific time range. It’s labeled All time by default. This is useful for seeing what an asset’s dependencies looked like as of a past date, or for narrowing a large graph down to recent activity. Choose one of the following:

• All time: Show edges from any time (the default).
• Last 7 days / Last 14 days / Last 28 days: Show only edges observed in that rolling window.
• Custom Range: Pick a specific start and end date.
• Point in time: Show the graph as it looked on a specific date, with no start date.

A time window is a hard prune during graph traversal, not a filter applied after the fact. If an edge falls outside the selected window, everything reachable only through that edge is removed from the graph too, even if those further edges are themselves inside the window. A node appears only if every edge on the path from your starting asset to it is within the window. Edges with no timestamp, from older or legacy lineage data, always match any window. This preserves backwards compatibility. A window over a mix of legacy and timestamped edges returns all legacy edges plus whichever timestamped ones fall inside the window, not a clean point-in-time snapshot. The time window applies to both the Lineage graph and the Impact Analysis table below.

​

Impact Analysis View

The Lineage page has two tabs at the top: Lineage (the graph you’ve seen so far) and Impact Analysis, a table that lists the same upstream and downstream assets as rows instead of nodes on a graph. This is easier to scan, search, and sort when an asset has a large number of dependencies.

• Upstream / Downstream: Switch direction. Each option shows a count of how many assets lie in that direction.
• Impact On: Choose whether the table lists Asset Level impact (one row per table, pipeline, dashboard, and so on) or Column Level impact (one row per source-to-target column mapping).
• Customize: Choose which columns to show. Name (or Source/Impacted Column, in Column Level mode) always stays visible. Every other column can be toggled off.

Search narrows the rows by asset name (Asset Level) or column name (Column Level). Asset Level columns: Name, Node Depth (number of hops from the asset you’re viewing), Description, Domains, Owners, Tier, Tags, and Glossary Terms. Column Level columns: Source, Source Column, Impacted (the downstream or upstream asset), Impacted Column, plus the same Node Depth, Description, Domains, Owners, Tier, Tags, and Glossary Terms columns.

​

Filtering by Service and Asset Type

Use the filters above the Impact Analysis table to narrow a large result. The filters include Service, Service Type, Domains, Data Products, Owners, Tag, Tier, Database, Schema, and Column. At Column Level, Glossary Term is also available.

• Inclusion only: Each filter keeps the values you select. Filters don’t support excluding a value. To hide a service, select every service you want to keep.
• Services map to one asset type: A service holds one kind of asset. For example, a Snowflake or Glue service holds tables, and an Airflow service holds pipelines. To list only pipelines, filter by the pipeline services. To list only tables, filter by the database services.
• Filters combine: When you select values in more than one filter, Collate keeps only the assets that match all of them.

​

Data Asset Details

Click a data asset to view its details:

• Source, Name, Description, Owner (team/user), Tier, and Usage information.
• Additional details based on the type of data asset (Table, Topic, Dashboard, Pipeline, ML Model, Container). For example, a table’s type, number of queries, and columns.
• Data quality and profiler metrics, including tests passed, aborted, and failed.
• Tags associated with the data asset.
• Schema details: column names, column types, and column descriptions.

​

Column-Level Lineage

Click a table to expand its list of columns and see column-level lineage.

​

Pipeline and Dashboard Lineage

For Pipelines:

• Lineage first comes from the metadata ingested from the databases.
• Setting up pipeline ingestion with a database service name links the pipeline to the database tables it reads from and writes to.
• If a pipeline creates the lineage, that shows up in the edge information too.

Dashboards work the same way:

• Lineage first comes from the metadata ingested from the databases.
• Dashboard ingestion then adds the data models and charts, linking them back to the database tables they’re built on.

​

Lineage Layers

Switch between layers to explore different dimensions of your data’s lineage.

​

Column Layer

Trace individual fields — like customer_id or first_name — as they move across tables and pipelines. The column layer shows exactly how each attribute transforms along the way.

​

Observability Layer

See data quality results directly in the lineage graph — test passes, failures, and pending checks appear on each node. This helps you quickly spot which assets in the pipeline have quality issues.

​

Service Layer

See how data moves across services — from Hive to Redshift to Power BI to Tableau. The service layer gives you a system-level view of the full data journey, from ingestion through transformation to consumption.

​

Domain Layer

Group assets by business domain — like “Ecommerce” or “Customer Data” — to understand lineage in business terms. This view connects your technical assets to the business functions they support.

​

Data Product Layer

See the lineage of your curated, consumption-ready data products — like Customer Registry or Superstore. The data product layer shows which trusted outputs belong to which domain and where they come from.