Skip to main content
Platinur does not require all data to be loaded by Platinur. Customers may load source data into ClickHouse through their own ingestion processes. The product contract is simple: one source maps to one ClickHouse schema. Platinur can inspect those schemas, let the operator choose tables, and use selected tables during Initial Run and assistant-driven model changes.

​
When to use external schemas

  • The customer already has ingestion jobs.
  • Source loading is governed by another platform.
  • The source requires a connector that is not enabled in the Platinur worker image.
  • The customer wants to separate ingestion ownership from modeling ownership.

​
Discovery during Initial Run

Every non-system ClickHouse schema that is not a Platinur output schema appears automatically in the Initial Run source list, with its tables and approximate row counts — including schemas with custom names that were loaded entirely outside Platinur. Toggle a discovered schema on, optionally narrow the table selection, and run. Selecting it for a run saves it as a configured source; no integration setup is required.

​
Operator responsibility

Before Initial Run, confirm that source schemas exist, have usable tables, and contain data appropriate for analytics modeling. Then use Operations to select the source schemas and tables Platinur should model.