Database Management Systems

Partition by Round Robin distributes records evenly across partitions by cycling through partitions sequentially - first record to partition 1, second to partition 2, and so on, then back to partition 1. This ensures balanced distribution regardless of data characteristics. Key partitioning causes skew if keys aren't uniformly distributed, Expression partitioning depends on the expression logic, and Range partitioning creates ordered but potentially uneven partitions.

DMX (Data Mining Extensions) uses SELECT FROM syntax for browsing and prediction queries against mining models. This allows you to retrieve patterns and make predictions from trained models. The other options like INSERT INTO, CREATE MINING STRUCTURE are for different operations (training, creation).

Relational OLAP (ROLAP) stores data in relational databases and is most suitable when you have few dimensions and small dense datasets. ROLAP avoids the preprocessing overhead of MOLAP and works well when data doesn't have significant sparsity. For many dimensions or large datasets, MOLAP or HOLAP would be more appropriate.

A dependent data mart is derived from and dependent on the enterprise data warehouse. It contains a subset of the data warehouse focused on a specific business unit or function. Independent data marts pull directly from source systems, but dependent ones are subsets of the centralized warehouse.

In a data warehouse, indexing should be applied to columns and data fields that are frequently referenced by queries to speed up data retrieval performance. Indexing every field or only fields changing over time would be inefficient and impractical.

Transaction activity records individual business events continuously in real-time, making it the most dynamic and frequently changing type of data. Product catalogs, organizational staff, and structural data are structural and change much less frequently.

Data mart design must begin with understanding the business requirements of the department that will use it. This determines what data to include, how to structure it, what metrics to calculate, and how to present the information. Database backups, query performance, data entry processes, and table counts are implementation or operational details that come after defining requirements.

The loadjava utility loads Java source code, class files, and resources into an Oracle database. When given a .java file, it compiles and loads the Java code into the database schema. It does not publish methods (that requires separate publication) nor does it specifically load metadata only into Java pool.

RESULT_CACHE_MODE=FORCE forces all query results to be cached if possible, ignoring the /*+ no_result_cache */ hint. The FORCE setting means caching happens regardless of explicit hints. Results are cached when they fit in available cache space, but hints that disable caching are overridden.

Internal LOBs can be used as attributes within user-defined data types (object types), allowing structured data with large object components. They also support separate tablespace storage from their parent table for better management. They DO use redo logging (contrary to option B) and CAN be passed as parameters to PL/SQL subprograms (contrary to option C).

DBMS_METADATA package provides GET_DEPENDENT_DDL to retrieve DDL for all objects that depend on a specified object (like a table). GET_DDL can fetch DDL for objects of a specific type within a schema, such as all tables. It doesn't directly track object invalidation status (B is false) or provide tablespace-level object listings (D is false).

When an Application Engine program is loaded into memory in PeopleSoft, the physical tables that the program will operate on are locked to prevent concurrent modifications. This ensures data integrity during program execution and prevents conflicts from other processes.