Database Management Systems

In SAP BW, when an InfoCube is compressed, the Request IDs (request IDs that track data loads) are deleted. The data is rolled up from the request-specific level to a more aggregated form. Dimension IDs, Surrogate IDs, and Master Data IDs are preserved.

Conformed dimensions are dimension tables that share the same structure and meaning across multiple star schemas or data marts. This allows queries to be executed across different schemas seamlessly. A Calendar dimension is a classic example - it can be used consistently across sales, inventory, and financial data marts.

A Data Warehouse is subject-oriented, integrated, time-variant, and non-volatile. 'Process Oriented' is characteristic of operational systems (OLTP), not data warehouses.

Data warehouses are highly DENORMALIZED, not normalized, for better query performance. Normalization (used in OLTP systems) reduces redundancy but requires many joins, which slows queries. Denormalization (used in data warehouses) increases redundancy but reduces joins, enabling faster analytical queries.

Data cleansing occurs primarily during the Transform phase of ETL (Extract, Transform, Load). During transformation, data is cleaned, validated, standardized, and corrected before being loaded into the data warehouse. The Extract phase only retrieves data, and the Load phase writes already-cleaned data to the target system.

OLAP (Online Analytical Processing) tools are specifically designed for creating reports and performing multidimensional analysis on data warehouses. OLTP systems are for transaction processing, and ETL tools are for data extraction/transformation/loading. The question asks about reporting tools, and OLAP is the correct category.

This statement is incorrect. A data warehouse can and often does have multiple fact tables. Different business processes (sales, inventory, financial transactions) are typically represented by separate fact tables in the same data warehouse. These fact tables share common dimension tables but represent different measurable events.

Star and Snowflake are both standard dimensional modeling schemas used in data warehousing. Star schema has a central fact table directly connected to dimension tables, while Snowflake normalizes dimensions into multiple related tables. Both are valid approaches depending on the trade-off between query performance and storage normalization.

The UNDOSTAT view in Oracle provides statistics on undo consumption, including undo block consumption rates. This is the primary view to analyze when sizing undo tablespaces. It helps estimate how much undo space your workload requires based on historical usage patterns.

If Automatic Undo Management is enabled but no undo tablespace is specified, Oracle will automatically select and use the first available undo tablespace in the database.