This set of Apache Hive MCQs covers advanced concepts of Apache Hive, including HiveServer2, SerDe, dynamic partitioning, query optimization, vectorization, skew joins, and operational mechanics of Hive. Useful for GATE, UGC NET, University Semester Exams, and Big Data Technology preparation.
Topic: Big Data Technology (Apache Hive) | Set: 2
Difficulty: Medium to Hard | Total Questions: 15
Important Apache Hive MCQs with Answers
Q1. Which component allows multiple clients to connect to Hive via JDBC/ODBC?
A. Hive CLI
B. HiveServer2
C. The Metastore
D. The JobTracker
View Answer & Explanation
Answer: B
Explanation: HiveServer2 enables remote clients to execute queries and retrieve results, supporting multi-user concurrency and authentication.
Q2. What is the function of “SerDe” in Hive?
A. Serializer/Deserializer; it tells Hive how to read/write data from/to files
B. Security/Dependency; it manages user access
C. Server/Deployment; it manages cluster installation
D. System/Error; it manages log files
View Answer & Explanation
Answer: A
Explanation: A SerDe allows Hive to process data formats like CSV, JSON, or Regex by defining how to serialize data into rows and deserialize it back.
Q3. Which Hive setting is used to enable dynamic partitioning?
A. hive.exec.dynamic.partition=true
B. hive.partition.dynamic=on
C. hive.dynamic.partition=enabled
D. set hive.dynamic.partition=1
View Answer & Explanation
Answer: A
Explanation: This configuration property must be set to true to allow the creation of partitions at runtime based on column values.
Q4. What is the purpose of ANALYZE TABLE … COMPUTE STATISTICS?
A. To format the HDFS drive
B. To collect metadata for the cost-based optimizer (CBO) to generate better execution plans
C. To delete old rows
D. To increase the HDFS block size
View Answer & Explanation
Answer: B
Explanation: By gathering statistics, the optimizer can choose more efficient join strategies, such as Map-side joins instead of Shuffle-side joins.
Q5. In Hive, how do you perform a “Map-side Join”?
A. By ensuring one of the tables is small enough to fit in memory and using the /*+ MAPJOIN(table_name) */ hint
B. By partitioning all tables
C. By increasing the number of reducers
D. By using external tables
View Answer & Explanation
Answer: A
Explanation: A Map-side join avoids the expensive Shuffle phase by broadcasting a small table to all map nodes and performing the join locally.
Q6. Which property controls the number of reducers in Hive?
A. mapred.map.tasks
B. mapred.reduce.tasks
C. hive.exec.reducers.bytes.per.reducer
D. Both B and C
View Answer & Explanation
Answer: D
Explanation: While mapred.reduce.tasks can set it manually, hive.exec.reducers.bytes.per.reducer allows Hive to auto-calculate the reducer count based on data volume.
Q7. What is the “Vectorized Query Execution” feature?
A. A way to run queries on GPUs
B. A method to process a batch of 1024 rows at once instead of one row at a time to reduce CPU overhead
C. A security setting
D. A compression codec
View Answer & Explanation
Answer: B
Explanation: Vectorization improves performance by reducing function call overhead during query processing.
Q8. How does Hive handle a “Slow Node” (Straggler) during execution?
A. It cannot handle slow nodes
B. It uses Speculative Execution to launch duplicate tasks on other nodes
C. It pauses the whole cluster
D. It kills the job
View Answer & Explanation
Answer: B
Explanation: If a task takes too long, Hive (via MapReduce/Tez) can launch a speculative copy and use whichever finishes first.
Q9. Which file format is highly recommended for OLAP workloads in Hive?
A. TextFile
B. SequenceFile
C. ORC or Parquet
D. CSV
View Answer & Explanation
Answer: C
Explanation: ORC and Parquet are columnar formats that provide better compression and significantly faster query performance for analytical workloads.
Q10. What is the difference between LOCAL DATA INPATH and DATA INPATH in a LOAD DATA statement?
A. LOCAL loads from the local filesystem; the other loads from HDFS
B. LOCAL loads from HDFS; the other from the local filesystem
C. There is no difference
D. LOCAL is faster
View Answer & Explanation
Answer: A
Explanation: LOAD DATA LOCAL expects the file on the client’s machine; LOAD DATA expects it on HDFS, which Hive then moves into the table directory.
Q11. Which command is used to change the column name of an existing table?
A. ALTER TABLE name CHANGE COLUMN …
B. RENAME COLUMN …
C. UPDATE COLUMN …
D. MODIFY COLUMN …
View Answer & Explanation
Answer: A
Explanation: Hive uses the CHANGE COLUMN clause within the ALTER TABLE command to modify column names, types, or positions.
Q12. What is the “Hive Thrift Server”?
A. A service that provides programmatic access to Hive from languages like Python or Java
B. A server for HDFS backups
C. A web UI for YARN
D. A compression too
View Answer & Explanation
Answer: A
Explanation: The Thrift Server allows external applications to interact with Hive via the Thrift framework, often used with clients like PyHive.
Q13. What is “Skew Join” optimization?
A. A technique to rebalance joins when one key appears much more frequently than others
B. A way to join on dates
C. A method to rename keys
D. A way to sort keys
View Answer & Explanation
Answer: A
Explanation: Skew joins prevent reducers from crashing or bottlenecking by handling high-frequency keys differently than standard join keys.
Q14. What is the purpose of UDF in Hive?
A. User-Defined Function; to add custom logic not available in standard HiveQL
B. Universal Data File; a storage format
C. User-Defined Format; a compression codec
D. Unified Data Framework; a scheduler
View Answer & Explanation
Answer: A
Explanation: UDFs allow developers to write custom Java code to perform complex transformations that HiveQL functions cannot achieve.
Q15. How can you disable MapReduce and run Hive queries in local mode?
A. SET hive.exec.mode.local.auto=true;
B. SET mapred.job.tracker=local;
C. SET hive.execution.engine=local;
D. By stopping YARN
View Answer & Explanation
Answer: A
Explanation: Enabling this setting allows Hive to run small jobs on the client machine instead of submitting them to the YARN cluster.
Conclusion
These Apahce Hive MCQs questions covered advanced concepts such as HiveServer2, SerDe, dynamic partitioning, vectorized execution, skew joins, query optimization, and operational mechanics in Hive.
Practicing these questions is useful for Big Data Technology, Hadoop ecosystem learning, GATE CS, UGC NET, and university semester examinations.
For better understanding of theory and concepts, refer to Apache Hive