Data Flows: Building Reusable Data Pipelines in OAC

Explain how OAC Data Flows let users build reusable, multi-step data transformation and machine learning pipelines without writing ETL code.

Data Flows are OAC's visual, no-code (or low-code) pipeline builder — the natural next step up from one-off Data Preparation transformations, meant for repeatable, multi-step data processing that produces a new dataset other projects can consume. Interviewers raise Data Flows to test whether a candidate understands OAC's self-service ETL capability, distinct from both classic ODI-based warehouse ETL and simple ad-hoc DV blending.

A Data Flow is like a home assembly-line kit: you snap together pre-built stations (filter, join, calculate, predict) in whatever order you need, and once assembled, the same conveyor belt runs automatically every time new raw material arrives — without needing an engineer to redesign the factory floor each time.

Key Concepts

A Data Flow is built as a sequence of connected steps on a visual canvas, each step performing one operation: adding a data source, filtering rows, joining two sources, aggregating, adding calculated columns, pivoting/unpivoting, applying a machine learning transformation (train a model, or apply an already-trained model to score new data), or writing the final output to a new dataset (which can then be reused as input to other projects or even other Data Flows, enabling chained pipelines). Each step's output can be previewed at every stage, which makes debugging a broken pipeline much faster than troubleshooting a black-box query.

stepsmachine learning transformation

Data Flows can be run on demand or scheduled to execute automatically and refresh their output dataset — this is what lets a business analyst build something functionally similar to a lightweight ETL job (clean, join, aggregate, and machine-learn) entirely within OAC's UI, without touching ODI, without writing PL/SQL, and without needing a data engineer for every iteration. This democratization is one of OAC's core value propositions over classic OBIEE.

scheduled

A sharp interview answer notes where Data Flows *don't* replace real ETL: they're generally best suited for analytical, dataset-to-dataset transformations feeding OAC visualizations, not for high-volume, mission-critical production data warehouse loading with full auditability, error handling, and enterprise orchestration — that's still squarely ODI/enterprise-ETL territory, and Data Flows and ODI-based ETL typically coexist, serving different layers of the same organization's data pipeline.

analytical, dataset-to-dataset transformations