For everyone

Lineage and connections

Where data comes from and where it goes, and how to read the picture.

Open lineage

Open an asset. See Reading an asset.
Under Trace it, press Lineage and connections.

You can also press Lineage on the Best match card in Ask. Press ← back to the asset to return.

Read the diagram

Data flows from left to right, one column per step. Arrows show data moving one hop.
Rows are the layers of the model, such as business, logical and physical. Your repository's metamodel names the layers.
A dashed line between two rows is one thing seen at two levels, for example a data entity and the table that realises it. It is not data moving, and it does not count as a hop.
The asset you started from is highlighted.

Point at a line and its name appears in the toolbar above the diagram. Hover over a box to see its name and type. Click a box to open that asset's page.

Not on this lists related assets the diagram leaves out, such as data domains, business terms and containers. Owning, containing or defining something is not moving data, so those links are shown under Connections instead.

Under the diagram, a list shows every asset on it. Side says upstream, downstream or here. Hops says how far away each asset is, and Reached by says which link reached it. The list stays readable when the diagram gets too crowded.

Choose what the diagram shows

Use the controls above the diagram.

Control	What it does
View	Lineage shows data flow. Connections shows everything related, at any distance you choose, whatever the kind of link.
Direction	Both ways, Upstream only (where data comes from) or Downstream only (what it affects). Lineage only.
Grain	Column level, Object level (tables and views) or Business level (what things are called). Lineage only.
Hops	How many steps out to follow.

If More beyond appears, there is more lineage past the edge. Raise Hops to follow it. If This is dense. appears, lower Hops, collapse a layer or use the list below the diagram.

If Full appears, the picture holds as much as a diagram can show and there is more connected than that. Raising Hops will not add to it. Open the graph from something larger instead — a table or a schema rather than a column — and you get a picture of the part you meant.

At Business level, the diagram shows the same flow drawn between named things such as entities, products and metrics. Nobody has named lists assets that move data but have nothing at the business layer standing for them.

It names the twelve that carry the most data through this picture, and says and N more for the rest — not the first twelve alphabetically, which on a large repository meant a list of things beginning with A and never the view everybody was actually waiting for. The count beside the heading is of all of them, and the Excel export of this screen carries every one, so the list is a place to start rather than the whole job.

When an asset has no lineage but does have other links, the screen opens on Connections and says why.

Group assets by what holds them

Columns start folded into the table or view that holds them. A box shows how many it holds, for example 7 Columns.

Click the count on a box to show those columns on the diagram.
The Inside row lists every box that holds something. Press a name to open or fold that one.
Press Show all to draw every column, or Fold … away to put them all back.

A line to a folded column is drawn to its table instead. Nothing is lost: every asset is still in the list under the diagram.

Choose how far out to group

The Grouped by row has a picker for each layer, and each one offers the steps of nesting that layer actually has. For the physical estate that is usually:

Not grouped · Table or View · Schema · Database · Source System

Pick Schema and every column and table in the picture is drawn as the schema that holds it. Pick Source System and a whole warehouse becomes one box with the count on it. It is the fastest way to turn "sixty boxes I cannot read" into "these two systems, and the line between them".

Each layer is chosen separately, because grouping the physical estate by schema while leaving the logical model alone is the ordinary request. Not grouped draws every asset as its own box.

The steps come from your own metamodel — whatever the workbook nests inside what — so a repository that puts fields inside records inside files gets those steps instead, with no setting to change. A layer whose assets are not nested inside anything has no picker, because there is nothing to choose.

Grouping never loses anything. Every asset is still in the list under the diagram, one click on a box opens it, and the exported workbook has a row for every asset saying which box it went into.

Collapse layers

The Layers row has a chip for each layer, with a count.

Press a layer's chip to collapse the whole layer.
Press ▸ beside a chip to collapse single kinds of asset inside the layer instead.
Use New collapses to choose what the next collapse does: Fold into one box, Fold — click the box to open it, or Hide the layer entirely.

Collapse all but this one folds every layer except the one your asset is in. Show everything brings them all back. Notes under the diagram say what was Folded and what was Hidden.

Save a way of reading it

A saved lineage view keeps your choice of view, direction, grain, hops, collapsed layers and opened tables.

Set the diagram up the way you want it.
Press Save this view…, and give it a name.
Choose where it applies: Just this asset, Every asset of this type, or Everywhere — my default.
Press Save.

A view saved on one asset takes priority over one saved on its type, which takes priority over one saved everywhere. The matching view opens by itself. Pick another one from Saved view. If you change a control, changed since it was saved appears. To keep the change, press Save this view… and then Update with the view's name.

A saved view starts private. You can share it when you save it, under Can see it, or later with Share… on a view you saved. Choose everyone, a role, a team or a person. A role share reaches whoever holds that role over the asset, including people appointed later. Copy link copies a link that opens this asset with the view applied.

Note: Saving a view needs VIEW_CREATE, and sharing one needs VIEW_SHARE. If you do not have them, the screen tells you instead of showing the button.

Export

Press Export in the toolbar above the diagram.

The picture, as you see it: PNG, JPEG or PDF.
The data behind it: an Excel workbook, or CSV files for the assets, the links and what is left out. These are available in the Lineage view only.

Collapsing changes the picture, not the file. Every asset is still a row, marked with what the picture did to it.

The toolbar also has zoom buttons, Fit, 100% and Full page. Press Esc to leave full page.

Explore the whole graph

Open the whole graph in a new tab (on the lineage screen), or whole graph in a new tab (on an asset page), opens everything connected to the asset in a free layout.

Drag a node to move it. It stays where you put it. Drag the background to pan, and scroll to zoom.
Click a node to see its links in the side panel. Click a link to move along it.
A node that contains other assets has a small badge. Click the badge, or double-click the node, to put its contents away. The badge then shows how many are hidden. Click it again to bring them back. Put contents away and Open everything do this for the whole graph.
Find an asset… jumps to a node by name. Hops limits how far out to go, and Labels shows or hides names.

The line above the graph says how many assets are connected, how many are on the diagram and how many are collapsed.

On a large repository almost everything is connected to almost everything else, so the graph is read as far out as it fits rather than to the end. When that happens the line says the nearest N assets — more are connected, and Hops offers As far as it fits instead of All. It is not a fault and nothing is hidden from you: a diagram of a million boxes is not a diagram. Open it from a schema or a table to see the part you are asking about.