CDP-3002 無料問題集「Cloudera CDP Data Engineer - Certification」

（A）Both B and C

（B）RDDs are mutable, while DataFrames are immutable

（C）RDDs represent distributed datasets, while DataFrames work with structured data

（D）RDDs are lower-level abstractions, while DataFrames offer higher-level functionalities

（A）When querying data using the exact partition key in the WHERE clause

（B）When the partitioned table contains a small number of partitions

（C）When the table is partitioned on a column frequently used in query filters

（D）When querying data using a non-partition column as a filter

（A）'coalesce' can increase the number of partitions without shuffling.

（B）'coalesce' triggers a full shuffle of the data, improving data distribution.

（C）'coalesce' reduces the number of partitions without a full data shuffle, enhancing performance.

（D）'repartition' is incapable of reducing the number of partitions.

（A）To determine the number of partitions created during shuffle operations

（B）To directly control the amount of memory available to each executor

（C）To limit the maximum size of Spark's internal data structures

（D）To specify the number of tasks that can be run in parallel on a single executor

（A）'spark.executor.memory'

（B）'spark.driver.memory'

（C）'spark.executor.instances'

（D）'spark.executor.memoryoverhead'

（A）Wide tables require more storage and may lead to performance overhead due to data serialization.

（B）Wide tables offer better performance for filtering operations, while narrow tables excel in aggregations.

（C）Both A and C.

（D）Spark can optimize data access patterns for both wide and narrow tables equally.

（A）They decrease the security of the data.

（B）They automatically encrypt data based on the partition key.

（C）They ensure that all data is stored in a single file, simplifying access.

（D）They allow for faster query processing by scanning only relevant partitions.

（A）Lineage tracks the dependencies between data transformations, enabling efficient re-execution on failures.

（B）Lineage allows for caching intermediate results, improving performance.

（C）Both A and B.

（D）It creates a log of all operations performed on data, aiding in debugging issues.