Database Management Systems

With a Partitioned Primary Index (PPI), data distribution across AMPs is still determined by the Primary Index hash value, just like a regular PI. The partitioning component adds a secondary organization layer within each AMP, but does not affect how rows are distributed to AMPs initially. The partitioning column defines how data is organized within each AMP, not across AMPs.

PPI provides two key advantages: (C) It enables efficient range queries and eliminates full table scans without requiring a secondary index structure, and (D) It ensures all data within a specific range is stored on the same AMP, improving range query performance. Option A is incorrect because NUPI, not PPI, guarantees uniqueness. Option B is incorrect - while PPI facilitates data removal/addition by partition, it's not instantaneous.

A transaction is defined as a logical unit of work (LUW) - one or more SQL statements that are treated as a single atomic operation. Either all statements in the transaction complete successfully, or none do (ACID properties). A transaction is not simply a logon/logoff sequence, nor is it limited to a single SQL statement, and while it relates to recovery, it's not defined solely by recovery units.

Normalization is the process of organizing data in a database to minimize data redundancy and dependency. Views, aliases, and sequences do not reduce redundancy; they serve for abstraction, renaming, and generating unique numbers.

Default logging in performance testing tools like LoadRunner is minimal - only errors are logged. This reduces I/O overhead during load tests, which is critical for accurate performance measurements.

Drill-through reports enable users to click on a data point in one report and navigate to a related report with more detailed or contextual information about that specific value.

The Informatica Repository stores ALL metadata including transformation rules, mappings, workflows, source definitions, target definitions, and sessions. However, given the question structure requiring a single choice, 'Metadata for transformation' is the most comprehensive and central answer. Source and target schemas are stored but are subsets of the broader metadata. The repository is essentially a metadata database managing all ETL artifacts.

Informatica PowerCenter Server uses native database drivers for optimal connectivity and performance. For Oracle databases, it uses Oracle Call Interface (OCI) client libraries rather than generic ODBC or JDBC drivers.

Universe is the semantic layer in BusinessObjects that maps database objects to business terminology. It acts as an abstraction layer between the physical database and end-user reporting tools like WebI and DeskI. The universe contains classes, objects, and measures that users see in their reporting interface.

Return ports are available in Lookup and Stored Procedure transformations to pass values back to the mapping. Filter and Rank transformations do not have return ports.

Stored Procedure transformations can be Passive, Connected, or Unconnected. They are not classified as Active because they execute procedures directly in the database.

When Informatica server cannot write to required directories (session log, bad files, cache files), the session initiates but fails during execution when it attempts to write these files. The initialization succeeds, but the runtime fails due to permission issues.

RDF stands for Record Definition Field in database/records management contexts. The other options (Relative Data Field, Relative Definition Field, Record Data Field) are not standard abbreviations. RDF is used to define the structure and format of data records.

In IBM Tivoli Identity Manager (TIM) CSV feed configuration, the name attribute maps to the attribute that uniquely identifies each individual person record in the source CSV file during reconciliation.

Excluding attributes unrelated to policy evaluation reduces data transfer and processing overhead, improving reconciliation performance. Weekly reconciliation frequency is unrelated to performance optimization. Disabling policy evaluation would compromise functionality, and JVM tuning should follow proper sizing rather than arbitrary increases.