Database Management Systems

Data warehouses and data marts both use denormalized schemas to optimize query performance. Denormalization reduces joins and improves read speed, which is essential for analytical processing. Both can exist in the same tier of an architecture (e.g., an ODS feeding both a DW and multiple marts). Option D correctly states this.

Report purge in BusinessObjects refers to the deletion of data retrieved by the query. When a report is purged, the data rows are removed but the report structure (layout, formatting, formulas) remains intact. This is different from deleting the entire report structure or refreshing it with new data. Option B correctly describes this.

In Web Intelligence multi-query documents, only Dimension objects can be merged across queries. Measures cannot be merged because they represent calculations that would conflict when combined. Detail objects are query-specific and cannot be merged. Dimensions provide the consistent reference points needed to combine data from different queries.

Derived tables reduce maintenance overhead because they eliminate the need to manually create and update physical database summary tables. They provide real-time data analysis by executing queries against current data rather than potentially stale pre-aggregated tables. Option A is incorrect because derived tables can include complex calculations and functions.

In OLAP (Online Analytical Processing), the standard output format is a matrix where dimensions form the rows and columns, and measures (the numerical values) populate the cell values. This is the fundamental structure of a pivot table or cube view. Option A incorrectly places 'Values' in the second position, and Option B misorders the components entirely.

Detail objects in Business Objects universe design are always attached to dimensions and provide additional descriptive information about those dimensions. They cannot exist as standalone objects (Option C) or attach to measures (Option A). This is fundamental - dimensions are the core objects, and details elaborate on them with extra context.

OLAP cache stores query navigation states and results in application server memory, enabling result sharing across users who execute the same query. Cache reads are faster than database reads since data is pre-computed and stored in memory. The cache size depends on server memory, not user desktop memory.

Aggregation provides OLAP with pre-calculated values, which are summary values computed in advance and stored. This enables faster query response times since calculations don't need to be performed at runtime.

Metrics are derived from Measure Objects, which contain numerical values that can be aggregated (sum, average, etc.). Dimension Objects provide descriptive context for grouping, while Detail Objects provide additional attributes about dimensions.

OLAP achieves multidimensionality through cube structures that organize data across multiple dimensions (time, geography, product, etc.). The cube structure enables complex and ad-hoc query analysis by allowing users to navigate and slice data along any dimension combination.

In COBOL ENVIRONMENT DIVISION, programs (not subroutines) should have both ERROR RECOVERY and ROLLBACK paragraphs. These paragraphs handle error conditions and transaction recovery. The question follows COBOL mainframe programming standards where recovery sections are required for robust program execution.

In certain enterprise COBOL programming schemas, the subroutine structure requires the ERROR RECOVERY paragraph in the ENVIRONMENT DIVISION to handle execution-time anomalies and provide structured recovery blocks, distinguishing it from standard programs which do not enforce this section.

A program (not a subroutine) in the ENVIRONMENT DIVISION should have both ERROR RECOVERY and ROLLBACK paragraphs (D). ERROR RECOVERY handles exceptional conditions, while ROLLBACK manages transaction rollback on failure. Having both ensures comprehensive error handling and data integrity. Either one alone is insufficient.

SQL Server 2000 was codenamed Shiloh during its development. Yukon was the codename for SQL Server 2005, Mantis was unrelated to SQL Server, and Aurum was not associated with any Microsoft SQL Server release.

Views are virtual tables that hide underlying table structure while presenting data in a customized way. They support abstraction and can be updatable, making them ideal for allowing controlled data access without exposing base tables.