Oracle is now an Artie destination

Until now, Oracle was source-only in Artie: you could replicate changes out of an Oracle database into a warehouse or another database, but you could not replicate data into one. Teams that run Oracle as a reporting database, a regional replica, or the system a legacy application reads from had no way to keep it fed from their other databases.

Oracle is now a destination as well. You point Artie at a pluggable database's service name and an existing schema, and it creates and evolves tables there as the source changes. Source types map to Oracle types (booleans to NUMBER(1, 0), floats to BINARY_DOUBLE, timestamps to TIMESTAMP(6), JSON and arrays to CLOB), and for Oracle-to-Oracle pipelines Reader now emits each column's exact source type, so NUMBER(10,2) or VARCHAR2(50 CHAR) arrives as the same type on the other side rather than a generic equivalent. Identifiers are uppercased the Oracle way, so a source column orderId becomes ORDERID.

The service account needs only CREATE SESSION, CREATE TABLE, and a tablespace quota to write into its own schema; writing into another schema needs the ANY TABLE privileges listed in the docs. Oracle 12.2 or later is required, and identity columns on pre-created tables should be GENERATED BY DEFAULT, since Artie writes the key.

This is for teams that need to keep an Oracle database in sync from PostgreSQL, MySQL, SQL Server, or another Oracle.

Why this matters

• One pipeline from a supported source into Oracle, set up from the Dashboard like any other destination
• Oracle-to-Oracle replication keeps your column types exact
• A minimal grant: the service account can be confined to its own schema