Search Results get_user_map

Overview

SYS.DBMS_HADOOP is a core Oracle Database-supplied PL/SQL package that exposes a set of Java-enabled entry points for integrating the database with Hadoop and Apache Hive environments. Within the Oracle E-Business Suite 12.1.1 and 12.2.2 technology stack, the package is installed automatically as part of the standard RDBMS build and is owned by the SYS schema, with a public synonym and EXECUTE grant that make it callable from any schema, including APPS.

The package's business purpose is to bridge the database and the Hadoop/Hive world. It allows EBS and DBA personnel to generate external table DDL that targets data stored in HDFS, and to obtain the mapping of Oracle database users to their corresponding Hive/OS users for authentication and impersonation scenarios. This is relevant for EBS deployments that offload reporting or large-volume interface processing to Hive, export EBS data extracts to Hadoop, or use Oracle Big Data SQL/Big Data Connector to query HDFS directly from SQL. ETRM 12.2.2 lists the package as VALID with an "OTHER" API classification, indicating it is infrastructure-level rather than a business API.

Key Procedures and Functions

The ETRM metadata documents two callable entry points for SYS.DBMS_HADOOP:

• CREATE_EXTDDL_FOR_HIVE — Produces the Data Definition Language required to create an Oracle external table over data that resides in a Hive table or HDFS location. It is used to materialise the mapping between a Hive schema and an Oracle external table so that the data becomes queryable from SQL without a physical data movement step.
• GET_USER_MAP — Returns the mapping between Oracle database users and their corresponding Hive or operating-system user identities. This mapping is consumed when the database must authenticate to Hive on behalf of a session, particularly where impersonation or delegated credentials are used.

No parameter lists are documented in the ETRM extract; parameter signatures should be confirmed against the database package specification before use in custom code.

Tables Accessed

The ETRM metadata for SYS.DBMS_HADOOP records no directly referenced application tables and no APPS-synonym dependencies. This is expected for a database-resident infrastructure package: it does not read or write EBS transactional tables. Any persistent configuration it relies upon is held in RDBMS-owned data dictionary views or in Java/Directory objects maintained by the database, not in EBS application schema objects.

Usage Notes

DBMS_HADOOP is not invoked by standard Oracle EBS forms, concurrent programs, or Oracle Application Object Library components. It is called from custom PL/SQL, database jobs, or ad hoc SQL sessions, and is dependent on the Java Virtual Machine being enabled in the database, since its implementation relies on the Oracle JServer (Java) infrastructure. Typical invocations include DBA or integration scripts that generate external table DDL over Hive datasets, and authentication routines in EBS-to-Hadoop connectors that must resolve Oracle users to Hive users.

Because the package is owned by SYS and its dependencies resolve through STANDARD and other SYS objects, it sits outside the normal EBS patching and AD utilities lifecycle. It is not customised, and changes are applied only through RDBMS patches or upgrades. Practitioners should verify the package specification for parameter names and datatypes at their specific database version before relying on it in production custom code, and should confirm that the RDBMS option and Java components are installed and valid. Status should be confirmed as VALID for each EBS 12.1.1 and 12.2.2 database, since an INVALID state will prevent execution.