The Flow service supplies data to VantageCloud Lake for initial loads, continuous loads, and ad hoc data exploration. You can schedule flows.
You can use Flow from the VantageCloud Lake Console.
A flow is a process that performs autonomous and continuous data loading and recovers from network and database failures. Flows run on the primary cluster. Flow guarantees that data is loaded exactly once.
Source Data
Source data for a flow comes from one or more object storage paths. A source is a path in an external object storage containing files. Supported file formats are CSV and Parquet.
You cannot join data from multiple sources in a flow.
Target Tables
If you do not specify an existing target table, Flow creates a target table, inferring the schema from the source data (see Schema Inference).
If you do specify an existing target table, Teradata recommends using schema hints to override the schema that Flow infers from the source data and help prevent load errors (see Schema Hints).
Source-Target Relationships
A flow can have up to five sources, but each source can have only one target. However, multiple sources in a flow can reference the same target table. For example, sources A, B, and C can have one target table, D.