The term “Big Data” is often characterized by:3Vs (Volume, Velocity, Variety)2Vs (Volume, Variance)4Cs (Capacity, Cost, Control, Coverage)NoneA) 3VsBig Data is defined by large size, fast processing, and diverse formats.
MapReduce works in two phases:Shuffle and ReduceMap and ReduceMap and CombineCombine and ShuffleB) Map and ReduceMap transforms data, Reduce aggregates results.
Apache Hive is primarily used for:Transactional processingData warehousingFile system managementMachine learningB) Data warehousingHive provides SQL-like queries on Hadoop.
Spark improves performance over Hadoop by:Using SQLIn-memory processingDisk-only processingLimiting nodesB) In-memory processingSpark stores data in RAM, reducing disk I/O.
HDFS ensures fault tolerance by:CompressionData replicationParallel computingShardingB) Data replicationData blocks are replicated across nodes to prevent loss.
HDFS stands for:High Distributed File SystemHadoop Distributed File SystemHybrid Data File StorageNoneB) Hadoop Distributed File SystemHDFS stores massive files across distributed nodes in Hadoop.
Hadoop’s processing framework is based on:MapReduceSparkPigHiveA) MapReduceMapReduce splits jobs into map and reduce tasks for parallel processing.
Which component manages Hadoop cluster resources?HDFSYARNNameNodeDataNodeB) YARNYARN allocates resources and manages job scheduling in Hadoop.
In Hadoop, large files are split into:TablesBlocksFramesPacketsB) BlocksFiles are divided into fixed-size blocks for distributed storage.
Which tool is used for querying structured data in Hadoop?PigHiveHBaseSqoopB) HiveHive provides SQL-like queries for Hadoop data.
Which Hadoop node stores metadata?DataNodeNameNodeTaskTrackerJobTrackerB) NameNodeNameNode manages metadata about file locations in HDFS.
MapReduce “map” phase outputs:Keys onlyKey-value pairsValues onlyBlocksB) Key-value pairsMapper produces key-value pairs processed by reducers.
Apache Pig scripts are written in:SQLPig LatinJavaPythonB) Pig LatinPig uses Pig Latin language for data analysis.
Spark is faster than Hadoop MapReduce because:It uses in-memory processingIt stores only in HDFSIt lacks schedulingNoneA) It uses in-memory processingSpark processes data in memory, reducing disk I/O.
Sqoop is used for:Streaming dataTransferring data between Hadoop and RDBMSFile compressionData visualizationB) Transferring data between Hadoop and RDBMSSqoop efficiently imports and exports data between Hadoop and relational databases.
Hadoop Distributed File System (HDFS) stores data in:RowsBlocksTablesRecordsB) BlocksHDFS splits files into fixed-size blocks stored across nodes.
MapReduce consists of two main tasks:Map and FilterMap and ReduceShuffle and SortRead and WriteB) Map and ReduceMap produces key-value pairs; Reduce aggregates them.
Which NoSQL database integrates with Hadoop?OracleHBaseMySQLDB2B) HBaseHBase is a columnar NoSQL store for large-scale data.
Hadoop’s fault tolerance relies on:ReplicationEncryptionShardingIndexingA) ReplicationHDFS replicates data blocks across nodes for reliability.
YARN in Hadoop primarily manages:Data blocksJob scheduling and resource allocationSQL queriesFile replicationB) Job scheduling and resource allocationYARN is Hadoop’s resource manager.
What is the default block size in HDFS (modern versions)?16 MB24 MB128 MB1 GBC) 128 MBHDFS uses large block sizes to handle big files efficiently.
In Hadoop, a combiner is used to:Merge HDFS blocksReduce network congestion during MapReduceCompress logsManage file metadataB) Reduce network congestion during MapReduceCombiners perform local aggregation before shuffling.
Which tool is best for real-time processing over Hadoop?HivePigSpark StreamingOozieC) Spark StreamingSpark handles real-time data better than batch MapReduce.
Which compression format is splittable in Hadoop?GzipBzip2ZipLZ4B) Bzip2Splittable formats allow parallel reading in clusters.
Hadoop uses which model for data processing?Client-serverMapReduceRPCDistributed cachingB) MapReduceMapReduce divides data into tasks that can be processed in parallel.
In Hadoop, "NameNode" is responsible for:Data storageManaging file system metadataProcessing dataNetwork routingB) Managing file system metadataNameNode tracks where data blocks are stored.
Hive is mainly used for:Data warehousing and SQL queries on HadoopFile storageData encryptionStreamingA) Data warehousing and SQL queries on HadoopHive translates SQL queries into MapReduce jobs.
Which file format is most efficient for storing structured data in Hadoop?CSVParquetTXTJSONB) ParquetParquet’s columnar storage improves performance and compression.
Spark’s RDD provides:Mutable collectionsImmutable distributed datasetsSQL interfaceMetadataB) Immutable distributed datasetsRDDs allow fault-tolerant parallel processing.
Hadoop Streaming lets you:Use only JavaRun MapReduce in any languageStream videoEncrypt dataB) Run MapReduce in any languageIt accepts scripts as mapper/reducer.
HDFS replication factor default is:1234C) 3Each block is stored on 3 nodes for fault tolerance.
Hive QL is similar to:PythonSQLC++RB) SQLIt provides SQL-like interface over Hadoop.
Flume is used for:ETL of log dataCluster managementSecurityCompressionA) ETL of log dataFlume moves streaming data to HDFS.
In Spark, lazy evaluation means:Results are computed immediatelyTransformations are delayed until an actionAll data is cachedMemory is pre-allocatedB) Transformations are delayed until an actionSpark optimizes execution before running transformations.
Apache Flink differs from Spark because it:Is sloweSupports real-time streaming nativelyLacks fault toleranceUses Hadoop onlyB) Supports real-time streaming nativelyFlink handles event-time streams with minimal latency.
Hadoop uses which system for distributed storage?HiveHDFSMapReduceSqoopB) HDFSHadoop Distributed File System stores large data across clusters.
What is the main function of MapReduce?StorageQueryingData processingVisualizationC) Data processingMapReduce processes large-scale data in parallel across nodes.
Which Hadoop component provides SQL-like querying?PigHiveHBaseFlumeB) HiveHive converts SQL queries into MapReduce jobs.
What kind of data can Hadoop handle?Structured onlyUnstructured onlyBoth structured and unstructuredEncrypted data onlyC) Both structured and unstructuredHadoop is designed for all types of massive data sets.
Which tool transfers data between Hadoop and relational databases?OozieSqoopFlumeSparkB) SqoopSqoop is used for efficient import/export of data between Hadoop and RDBMS.
MapReduce divides a task into which phases?Shuffle & ReduceMap & ReduceMap & CombineSort & FilterB) Map & ReduceMap handles data processing, Reduce aggregates the results.
Default HDFS block size in Hadoop 2.x is:64 MB128 MB 256 MB512 MBB) 128 MBHadoop 2.x uses 128 MB as the default storage block size.
Which tool transfers data between RDBMS and Hadoop?OozieFlume SqoopSparkC) SqoopSqoop imports and exports structured data between databases and HDFS.
MapReduce consists of:Sort and FilterSplit and Merge Map and Reduce phasesEncode and DecodeC) Map and Reduce phasesIt processes data in two main stages: mapping and reducing.
HDFS stores data by dividing it into:Segments BlocksChunks PagesB) BlocksFiles are split into blocks for distributed and parallel storage.
Which component of Hadoop executes jobs?HiveYARN SqoopMapReduceD) MapReduceMapReduce executes distributed tasks over Hadoop clusters.
Hive queries are written in:Pig LatinSQLHiveQLJSONC) HiveQLHive uses a SQL-like query language for data processing on HDFS.
In Hadoop, secondary NameNode is responsible for:Backup storage Log managementCheckpointing metadataScheduling tasksC) Checkpointing metadataIt merges edits and updates of the NameNode periodically.
Which Hadoop component converts jobs into smaller tasks to run in parallel?HDFSYARNMapReduceHiveC) MapReduceMapReduce divides a task into smaller chunks and processes them concurrently across nodes.
What is the main purpose of Hadoop’s NameNode?To store data blocksTo manage metadataTo replicate dataTo handle client requestsB) To manage metadataThe NameNode keeps track of block locations and file structure in HDFS.
What’s the role of a DataNode in HDFS?Managing job executionStoring actual data blocksMonitoring cluster health Replicating metadataB) Storing actual data blocksDataNodes physically store data blocks assigned by the NameNode.
Apache Pig is mainly used for:Data visualizationScripting large data transformationsDatabase designSecurity analysisB) Scripting large data transformationsPig provides a high-level scripting language to simplify big data analysis over Hadoop.
Spark is often preferred over Hadoop MapReduce because:It runs only on small data It supports in-memory processingt’s less scalable It uses relational databasesB) It supports in-memory processingSpark performs computations in memory, making it significantly faster than disk-based MapReduce.
What is MapReduce used for?Data encryptionParallel data processingQuery optimizationData compressionB) Parallel data processingMapReduce divides tasks into smaller map and reduce jobs for distributed computing.
Which component manages job scheduling in Hadoop?NameNodeJobTrackerDataNodeTaskManagerB) JobTrackerJobTracker coordinates MapReduce jobs across nodes.
What is the function of the NameNode?Executes tasksManages metadataStores dataCompresses dataB) Manages metadataThe NameNode tracks file locations and directory structure in Hadoop.
Which component handles job scheduling in Hadoop?DataNodeJobTrackerReducerMapperB) JobTrackerIt coordinates and manages all MapReduce jobs.
What is “data locality”?Storing metadata separatelyProcessing data near where it’s storedData encryptionDistributed cachingB) Processing data near where it’s storedReduces network traffic by executing tasks on nodes containing data.
What is YARN’s primary purpose?Manage resourcesEncrypt dataBackup filesSplit blocksA) Manage resourcesYARN coordinates resource allocation among cluster applications.
What does the NameNode in Hadoop do?Stores dataManages metadataExecutes map tasksHandles replicationB) Manages metadataThe NameNode tracks file locations and block information.
HDFS splits files into blocks of size:32MB64MB128MB256MBC) 128MBThe default block size is 128MB to optimize performance and fault tolerance.
Which component performs the actual data processing in Hadoop?DataNodeMapReduceHDFSYarnB) MapReduceMapReduce handles distributed data processing and computation.
In Hadoop, secondary NameNode is used for:BackupLoad balancingCheckpointingData compressionC) CheckpointingIt merges edit logs with the file system image for efficiency.
YARN stands for:Yet Another Resource NegotiatorYour Advanced Resource NetworkYield Allocation Resource NodeNoneA) Yet Another Resource NegotiatorYARN manages cluster resources and scheduling in Hadoop 2.0+.
MapReduce processes data in:One nodeParallel SequentiallyLocallyB) ParallelMapReduce distributes data tasks to multiple nodes for speed.
YARN in Hadoop is responsible for:Data storageResource management File readingDebuggingB) Resource managementYARN allocates computing resources dynamically to tasks.
Spark is faster than MapReduce because:It uses disk for computationIt uses in-memory processing It uses batch processingIt ignores failuresB) It uses in-memory processingSpark processes data in memory, reducing read/write delays.