How Data Catalog Links Work

The Data Catalog Links dialog shows how a Qualytics asset is linked to its counterpart in each connected data catalog, and lets you correct that link when the automatic match picked the wrong asset or found none. It works the same way across every connected catalog; for the list of catalogs Qualytics integrates with, see Supported Data Catalogs.

A link connects one Qualytics asset to one asset in one connected catalog. The same asset can hold a link in each catalog you have connected, so it may be linked in Alation and in Atlan at the same time, and each pair is managed on its own. This page covers where links come from, how manual and automatic links behave differently, and why the datastore's link matters more than the others.

The dialog exists on datastores, containers, and fields, each scoped to its own level. The datastore is the anchor of the feature. It is the only level that can start a synchronization from the dialog, and its manual link governs how everything under it resolves.

Availability

Data Catalog Links are available on datastores backed by a database connector (JDBC and native connectors), other than enrichment datastores. File-based (DFS) datastores do not offer the dialog, at any level. Enrichment datastores are left out because they hold the outputs Qualytics writes, which have no counterpart in your catalog. A connected data catalog integration is required.

Automatic Link

Most links are never touched by hand. During each catalog synchronization, the connector matches your assets to the catalog's by name and hierarchy. The datastore is matched to a schema-level asset, and its containers and fields are resolved underneath it. Each catalog has its own matching rules and supported asset types, documented in its setup guide.

A link created this way is an automatic link, and it stays under the synchronization's control:

• It is refreshed by later synchronizations, so it follows the catalog as assets move or are recreated.
• It is tied to the integration's configuration. When the configuration changes, existing automatic links are re-evaluated, and the next synchronization resolves them again.
• It never replaces a manual link. Where you have chosen an asset by hand, the matcher steps aside.

Automatic matching works well when names line up one to one. It runs into trouble when they do not, for example when two catalog sources hold tables with the same names, and that is where manual linking comes in.

Manual Link

When the automatic match picks the wrong asset or finds none, you can override it and point the asset at its true counterpart yourself.

A datastore links to a schema-level asset in the catalog, the level that represents a whole database or schema on the catalog's side.

You can find the counterpart by searching the catalog by name or identifier, or by pasting the asset's URL from the catalog. Qualytics extracts the identifier out of the URL and resolves the asset for you, and each catalog has its own address format.

Relinking to a different asset resets the link's Last Synced, so the next synchronization pushes the full Qualytics metadata to the newly linked asset instead of skipping it as unchanged.

Manual and Automatic Links

A link you set yourself is a manual link, and it behaves differently from the ones the synchronization creates:

• A manual link is never overwritten by a later synchronization. The automatic matcher leaves it alone.
• A manual link stays valid when the integration's configuration changes, while automatic links are re-evaluated.
• Unlink is only offered on manual links. Removing an automatic link would be pointless, since the next synchronization would just recreate it. After unlinking, the next synchronization is free to auto-match the asset again.

Multiple Catalogs

Links are kept per catalog. When a single data catalog is connected, the dialog opens straight on that catalog's link. When more than one is connected, the dialog shows one link card per catalog the datastore is linked in, and each link lives its own life:

• Each catalog gets its own link, created by its own synchronization or set by hand, so the datastore can be linked in Alation and in Atlan at the same time.
• Changing or unlinking the asset in one catalog leaves the links in the other catalogs untouched.
• A link can be manual in one catalog and automatic in another. The manual protections apply per link, not per asset.
• Synchronizations run per integration, so each catalog exchanges metadata with its own link on its own schedule.
• The limit runs in both layers. Each catalog can be connected once per deployment, and an asset holds at most one link per catalog, so linking another catalog always means a different one.

To add a link in a catalog the datastore is not linked in yet, see Link Another Catalog.

A Manual Datastore Link Anchors Its Children

Linking the datastore manually does more than fix that one asset. Later synchronizations use it as the anchor when resolving the datastore's containers and fields in the catalog. This is the way out when automatic matching cannot tell assets apart, for example when two catalog sources contain tables with the same names, or when the catalog search cannot reach child assets by name alone. Point the datastore at the right catalog source once, and the children resolve consistently under it.

Only the datastore link anchors. A manual link on a container or a field corrects that one asset and nothing else.

See Also