Skip to main content
Prizm connects to Oracle Database — on-premises, self-hosted VM, or Oracle Cloud Infrastructure (OCI), including Autonomous Database — to extract catalog metadata, observability signals, lineage, and quality measures using Oracle’s data dictionary and dynamic performance views. Supported on Oracle 19c and higher.

Why Connect Oracle to Prizm?

Connecting Oracle gives Prizm access to the same layers of intelligence available for other relational sources, adapted to Oracle’s data dictionary: Catalog & Context Prizm discovers every schema (Oracle user), table, view, column, stored procedure, and constraint, and indexes them in the Prizm catalog. Table and column descriptions are pulled directly from Oracle’s native COMMENTS — no re-entry required. Data Quality Oracle constraints that are already enabled — NOT NULL, primary/unique keys, foreign keys, and simple CHECK constraints — are automatically turned into activated quality rules on first catalog crawl. Auto-generated rules can be edited or disabled in Prizm without touching the underlying Oracle constraint. Observability Freshness, volume, and schema drift are computed from ALL_TAB_MODIFICATIONS, DBA_SEGMENTS, and consecutive catalog snapshots — no full table scans required for the core signals. Lineage Table-to-table lineage comes from foreign key constraints; view, procedure, and package lineage comes from ALL_DEPENDENCIES — both without SQL parsing for the standard case.

Authentication

Oracle connections authenticate with a Username and Password by default. Turning on Use Vault replaces those fields with a Select Vault dropdown — pointing at whichever vault instances your organization has already configured (AWS Vault, Azure Vault, or a named HashiCorp integration, for example) — plus a Key field for the credential’s path within that vault. See Setup for the full configuration walkthrough.

Criticality Scoring

Oracle tables and views are scored using the same weighted framework as other Prizm connectors, adapted to Oracle-specific signals: Tiers follow the same convention as SQL Server, Snowflake, and Databricks: Critical ≥ 80, High 60–79, Medium 40–59, Low < 40.
NUM_ROWS in ALL_TABLES reflects the last DBMS_STATS gather, not real-time state. Prizm treats volume scoring as stale when LAST_ANALYZED exceeds 7 days and falls back to DBA_SEGMENTS byte-based sizing.

Supported Oracle Objects

In Scope / Out of Scope

In scope: catalog, observability, and quality features across on-prem, self-hosted VM, and OCI Oracle deployments; authentication via username/password, Oracle Wallet, Kerberos, and enterprise vault integration; table/view/procedure/foreign-key lineage; criticality scoring; usage analytics from V$SQL, V$SQLAREA, and AWR (where licensed); catalog integrations with Alation, Atlan, Azure Purview, and Collibra; exception record extraction to Iceberg. Out of scope (this release): cost data, performance metrics beyond usage/session signals, and PL/SQL stored procedure execution.

Next Steps

Setup

Connect Oracle to Prizm — prerequisites, authentication, and configuration.

What We Collect

Full field-level breakdown of every metadata object Prizm extracts from Oracle.

FAQ

Common questions about the Oracle connector.

Troubleshooting

Diagnose connection, permission, and performance errors.