Analytic app data

An analytic-app-data node imports the result of a node from another of your analytic apps into this pipeline. It lets several apps build on one shared, already-published result instead of copying the same data or rebuilding the same steps in every app.

A common pattern is a shared "reference data" app whose output several reporting apps re-use, or a heavy data-preparation app whose result feeds a handful of smaller apps. Instead of duplicating that work, each downstream app imports the result with an analytic-app-data node.

Choose what to import

When you add an analytic-app-data node, you tell it what to bring in:

• the source analytic app to read from (one of your workspace's other analytic apps),
• the node inside that app whose result you want,
• optionally, which columns to import and in what order. Leave this empty to import every column exactly as it is.

The source-app list offers only the apps you may import from - apps whose data you can read, and an app in a personal space only for an app in that same space - and marks any app that has no published result yet, so you can only pick an app that is ready to import from.

How the data is captured

The data is imported as a snapshot of the source app's last successful published run, taken at the moment this pipeline runs. Two things follow from that:

• Only a published result is imported. A private preview run in the source app is never imported, so downstream apps only ever build on results their owners have deliberately published.
• The import is a point-in-time copy. When the source app publishes newer results, this app does not change on its own. Run this pipeline again to refresh the import with the newer values.

The imported node is read-only: it shows the imported result as a table, and you edit the data in the source app, not here. From this node the imported columns flow into the nodes connected after it, so you can blend the shared result with the rest of your data.

Who can change an app that imports another app

An analytic-app-data node carries another app's data into this one, so it is checked whenever this app's configuration changes:

• You can add an import node, or point one at a different app, only for an app whose data you can read. An app in somebody's personal space can be imported only into an app in that same personal space, and the Audit Analytics app can never be imported.
• Adding an import can therefore lock colleagues out. Before you save an import node that points at a new app, the node says how many people would no longer be able to modify this app, for example "3 people will no longer be able to modify this app". It never names them.
• Once this app imports from another app, changing anything in it - its pipeline, knowledge base, skills, Visualisation and settings, and publishing it - is open only to people who can read every app it imports from. Anybody else is told which import node stops them, and can ask for access to the app it imports from.
• They can still repair the app themselves, one import node at a time: deleting such a node, or pointing it at an app they can read, is always allowed. Nothing else is - not even choosing different columns of the same app. The same holds when the app an import node reads from has been deleted. A notice at the top of the editor names the import nodes that stop them, so they know which ones to change.
• Restoring an earlier version, or discarding a draft, is allowed only when the version it makes current imports from no app they cannot read; a version that does import from one stays out of reach, even if it imports from fewer of them.
• Running the pipeline, entering data, exporting and archiving are not affected.

Everybody who can read this app's data also sees what it imports, even without access to the source app - and only ever the source app's published result, never its drafts. Keep that in mind before importing from an app whose readers are a smaller group than this app's.