Spark API编程动手实战-02-以集群模式进行Spark API实战textFile、cache、count

操作HDFS:先要保证HDFS启动了:

启动spark集群:

以spark-shell运行在spark集群上:

查看下之前上传到HDFS上的”LICENSE.txt“文件:

用spark读取这个文件:

使用count统计该文件的行数:

我们可以看到count 耗时为0.239708s

对该RDD进行cache操作并执行count使得缓存生效:

执行count结果为:

此时耗时为0.21132s

再执行count操作:

此时耗时为0.029580s,这时因为我们自己基于cache后的数据进行操作的。

接着我们对上面的rdd进行wordcount操作:

通过saveAsTextFile把数据存到HDFS中:

我们通过web控制台查看下运行结果:

我们通过命令行看下part-00000和part-00001内容:

[[email protected] ~]$ hadoop fs -cat /data/resultLicenseWordCount/part-00000

15/01/22 13:51:32 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable

(under,10)

(Unless,3)

(Contributions),1)

(offer,1)

(agree,1)

(BUSINESS,2)

(NON-INFRINGEMENT,,1)

(its,4)

(materials,2)

(event,1)

(intentionally,2)

(Grant,2)

(writing,1)

(include,3)

(responsibility,,1)

(have,2)

(MERCHANTABILITY,,1)

(Contribution,3)

(Massachusetts,1)

(express,2)

("Your"),1)

((i),1)

(However,,1)

(been,2)

(files;,1)

(This,1)

(stating,1)

(2-Clause,1)

(conditions.,1)

(non-exclusive,,2)

(appropriateness,1)

(marked,1)

(risks,1)

(any,28)

(IS",4)

(implementation,1)

(filed.,1)

(Sections,1)

(fee,1)

(losses),,1)

(out,1)

(contract,2)

(DISTRIBUTION,1)

(4.,1)

(file,6)

(documentation,,2)

(wherever,1)

(unless,1)

(below).,1)

(names,,1)

(verbal,,1)

(ANY,10)

(version,1)

(file.,2)

(are,10)

(no-charge,,2)

(2.,1)

(from,,1)

(reproduction,,3)

(2011-2014,,1)

(assume,1)

(licenses,1)

(DATA,,2)

(IS,2)

(recommend,1)

(prominent,1)

(revisions,,1)

("[]",1)

(FITNESS,3)

(otherwise,,3)

(distribution,,1)

(necessarily,1)

(Apache,5)

(grant,1)

(CONTRIBUTORS,4)

(as,15)

(irrevocable,2)

(inclusion,2)

(purpose,2)

(products,1)

(ARE,2)

(merely,1)

(File,1)

(Definitions.,1)

(form,10)

(IMPLIED,4)

(Warranty,1)

(Patent,1)

(incurred,1)

(8.,1)

(repository,1)

(contributors,1)

("printed,1)

(sell,,2)

(:,3)

(malfunction,,1)

(Version,2)

(origin,1)

(alongside,1)

(CRC,1)

(implied.,1)

(contract,,1)

(representatives,,1)

(warranty,1)

(offer,,1)

(org.apache.hadoop.util.bloom.*,1)

(KIND,,2)

(is,10)

(conspicuously,1)

(found,1)

(charge,1)

(make,,1)

(file,,1)

(associated,1)

(even,1)

(same,1)

((Don‘t,1)

(outstanding,1)

(link,1)

([name,1)

(Trademarks.,1)

(notice,2)

(endorse,1)

(shall,15)

(contact,1)

(Redistributions,4)

(using,1)

(class,1)

(name),1)

(behalf,5)

(form.,1)

(We,1)

(INTERRUPTION),2)

(responsible,1)

(annotations,,1)

(THIS,4)

(subject,1)

(acting,1)

(permitted,2)

(OUT,2)

(BASIS,,2)

(has,2)

(Accepting,1)

(defend,,1)

(University,1)

([yyyy],1)

((http://www.one-lab.org),1)

(EVENT,2)

(granting,1)

(portions,1)

(implied,,1)

(NOTICE,5)

(infringed,1)

(limitation,,1)

(names,2)

(electronic,,1)

(PURPOSE,2)

(licensable,1)

(section),1)

(conditions,14)

(EVEN,2)

(acts),1)

(law,3)

(licenses.,1)

(compression,1)

(readable,1)

(solely,1)

(configuration,1)

(information.,1)

(litigation,2)

(represent,,1)

(warranty,,1)

(shares,,1)

(supersede,1)

(governed,1)

(marks,,1)

(http://code.google.com/p/lz4/,1)

(modification,,2)

(fifty,1)

(sent,1)

(places:,1)

(means,2)

(identifying,1)

(this,22)

(Works",1)

(Louvain,1)

(prior,1)

(slicing-by-8,1)

(PROCUREMENT,2)

(changed,1)

(describing,1)

(only,4)

(contributory,1)

(normally,1)

(indirect,,2)

(WITHOUT,2)

(Works,12)

(documentation,3)

(agreement,1)

(otherwise,3)

("AS,4)

(damages,,1)

(patent,,1)

(APACHE,1)

(without,6)

("NOTICE",1)

(Limitation,1)

(SUBSTITUTE,2)

(Contribution(s),3)

(Subject,2)

(Submission,1)

(UCL,1)

(TITLE,,1)

(trademarks,,1)

((iii),1)

(2.0,1)

(Fast,1)

(exercise,1)

(accepting,2)

(example,1)

(distribution.,2)

(interfaces,1)

(conditions:,1)

(act,1)

(incorporated,2)

(provides,2)

(limited,4)

(LZ4,3)

(2008,2009,2010,1)

(can,2)

(contents,1)

(PURPOSE.,1)

(recipients,1)

("Contribution",1)

(failure,1)

(communication,3)

(commercial,1)

(works,1)

(language,1)

(permissions,3)

(WARRANTIES,4)

(media,1)

(reserved.,2)

(Works,,2)

(How,1)

(WARRANTIES,,2)

(controlled,1)

(Warranty.,1)

(2.0,,1)

((http://www.opensource.org/licenses/bsd-license.php),1)

(own,4)

(submit,1)

(SHALL,2)

(reasonable,1)

(reason,1)

(agreed,3)

(systems,1)

(patent,5)

(form,,4)

(Technology.,1)

(advised,1)

(systems,,1)

(classes:,1)

(HOWEVER,2)

(distribution,3)

(DAMAGES,2)

((c),2)

(src/main/native/src/org/apache/hadoop/util:,1)

(PROFITS;,2)

(perpetual,,2)

(applies,1)

(apply,2)

(subcomponents,2)

(modify,2)

(owner],1)

(one,1)

(modifying,1)

(counterclaim,1)

(January,1)

(discussing,1)

(CONTRACT,,2)

(with,16)

((C),1)

(infringement,,1)

(2004,1)

(lawsuit),1)

(specific,2)

(LZ,1)

(warranties,1)

(reproducing,1)

(promote,1)

(beneficial,1)

(ADVISED,2)

((a),1)

(other,9)

(date,1)

(met:,2)

(publicly,2)

(from,4)

(LIMITED,4)

(display,,1)

(MERCHANTABILITY,2)

(damages,3)

(SUBCOMPONENTS:,1)

(negligence),,1)

(remain,1)

(CONDITIONS,4)

(their,2)

(electronic,1)

(identification,1)

(determining,1)

(consistent,1)

(display,1)

(writing,,3)

(trade,1)

(third-party,2)

(,1299)

(description,1)

(REPRODUCTION,,1)

(attached,1)

(list,4)

(*,34)

(INDIRECT,,2)

(designated,1)

(Contribution.",1)

(complies,1)

(addendum,1)

(damages.,1)

(Yann,1)

(EXPRESS,2)

(License;,1)

(6.,1)

(GOODS,2)

(subsequently,1)

(included,2)

(replaced,1)

(notice,,5)

[[email protected] ~]$   hadoop fs -cat /data/resultLicenseWordCount/part-00001

15/01/22 13:52:29 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable

(For,6)

(reproduce,,1)

("Contributor",1)

((or,3)

(nothing,1)

(work.,1)

(content,1)

(HOLDERS,2)

(add,2)

(through,1)

(All,2)

(perform,,1)

(result,1)

(goodwill,,1)

(herein,1)

(direct,,1)

(used,1)

(To,1)

(harmless,1)

(9.,1)

(these,1)

(control,,1)

(INCIDENTAL,,2)

(indicated,1)

(part,4)

(alone,1)

(different,1)

(forms,,2)

(purposes,4)

(https://groups.google.com/forum/#!forum/lz4c,1)

(be,7)

(/**,2)

(carry,1)

(separable,1)

(including,5)

(contained,1)

(combination,1)

(calculation,1)

(license,7)

(FOR,6)

(thereof,,2)

(ARISING,2)

(constitutes,1)

(but,5)

(types.,1)

(stated,2)

(archives.,1)

(obligations,,1)

(5.,1)

(Works;,3)

(nor,1)

("Legal,1)

(Work,20)

(whole,,2)

(Copyright,5)

(at,3)

(copyright,,1)

(Redistribution,2)

(object,1)

(copy,3)

(indemnify,,1)

(asserted,1)

(HADOOP,1)

(attach,1)

("control",1)

(support,,1)

("Object",1)

(give,1)

(THEORY,2)

(may,10)

(except,2)

("Work",1)

(sublicense,,1)

(IF,2)

(granted,2)

(project,2)

(authorized,2)

(SPECIAL,,2)

(BY,2)

(retain,2)

(or,65)

(transfer,1)

(fields,1)

(Licensor,,1)

((b),1)

((ii),1)

(2005,,1)

(of,75)

(does,1)

(transformation,1)

((INCLUDING,2)

(DIRECT,,2)

(management,1)

(modified,1)

(Licensed,1)

(percent,1)

(Header,1)

(original,2)

(Contributor,,1)

(native,1)

((INCLUDING,,2)

(PARTICULAR,3)

(limitations,1)

(THE,10)

(INCLUDING,,2)

(power,,1)

(CAUSED,2)

(de,1)

(appropriate,1)

(against,,1)

(TORT,2)

("Source",1)

(each,4)

(1.,1)

(following,10)

(Liability.,2)

(acceptance,1)

("You",1)

(sole,1)

(from),1)

(See,1)

(tracking,1)

(for,19)

(cause,2)

(alleging,1)

(obtain,1)

(reproduce,3)

(source,,1)

(control,2)

(EXEMPLARY,,2)

(TERMS,2)

(terms,8)

(syntax,1)

(SERVICES;,2)

(made,,1)

(BUT,4)

(compiled,1)

(issue,1)

("submitted",1)

(OneLab,1)

(algorithm,1)

(was,1)

(While,1)

(entity,,1)

(do,3)

(PROVIDED,2)

(no,2)

(License,10)

(entity,3)

(Contributions.,2)

(mean,10)

(individual,3)

(Institute,1)

(computer,1)

(notices,9)

(Neither,1)

(Licensor,8)

(STRICT,2)

(made,1)

(authorship,,2)

(bind,1)

((the,1)

(indemnity,,1)

(distribute,3)

(You,24)

(grants,2)

(brackets,1)

(meet,1)

(for,,1)

(service,1)

(in,31)

(trademark,,1)

(boilerplate,1)

(WAY,2)

(LOSS,2)

(distributed,3)

(LIABILITY,,4)

(submitted,2)

(public,1)

(OF,19)

(managed,1)

(derived,2)

(Source,8)

(use,,4)

(name,2)

(definition,,2)

(that,25)

(src/main/native/src/org/apache/hadoop/io/compress/lz4/{lz4.h,lz4.c,lz4hc.h,lz4hc.c},,1)

(customary,1)

(BSD,1)

(thereof,1)

(claims,2)

(CONSEQUENTIAL,2)

(translation,1)

(format.,1)

(construed,1)

(DAMAGE.,2)

(applicable,3)

(binary,4)

(regarding,1)

(European,1)

(excluding,3)

(END,1)

((d),1)

(choose,1)

(NO,2)

(BE,2)

(direct,2)

(retain,,1)

(modifications,,3)

(forum,1)

(owner,4)

(USE,2)

(informational,1)

(The,3)

(legal,1)

((50%),1)

(document.,1)

(received,1)

(such,17)

(institute,1)

(distribute,,2)

(WHETHER,2)

(page",1)

((except,1)

(loss,1)

(common,1)

(additions,1)

(BSD-style,1)

(Appendix,1)

(Use,1)

(disclaimer,2)

(resulting,1)

(ON,2)

(hereby,2)

(License.,11)

(software,3)

(whom,1)

(along,1)

(lists,,1)

(required,4)

(OR,18)

(ownership,2)

(SOFTWARE,2)

(the,122)

(includes,1)

(obligations,1)

(import,,1)

(not,11)

(either,2)

(terminate,1)

(if,4)

(stoppage,,1)

(provided,9)

(submitted.,1)

(all,3)

(permission.,1)

("License");,1)

(written,2)

(generated,2)

(consequential,1)

(Derivative,17)

(AND,11)

(rights,3)

(http://www.apache.org/licenses/,1)

(terms.,1)

(Catholique,1)

(deliberate,1)

(entity.,2)

(Work,,4)

(special,,1)

(Additional,1)

(Legal,3)

(034819,1)

(least,1)

(text,4)

(on,11)

(editorial,1)

(redistributing,2)

("License",1)

(against,1)

(permission,1)

(9,1)

(separate,2)

(and/or,3)

(LICENSE,1)

(union,1)

((and,1)

(1,1)

(including,,1)

(Entity,3)

(negligent,1)

(LIABLE,2)

(IN,6)

(use,8)

(enclosed,2)

(contains,1)

(files,1)

(Entity",1)

(Work.,1)

(owner.,1)

(preferred,1)

(modifications,3)

(brackets!),1)

(available,1)

(code,5)

(http://www.apache.org/licenses/LICENSE-2.0,1)

(more,1)

(possibility,1)

(product,1)

(liable,1)

(SUCH,2)

(direction,1)

(must,8)

(making,1)

(Disclaimer,1)

(disclaimer.,2)

(Commission,1)

(OTHERWISE),2)

(Hadoop,1)

((an,1)

(APPENDIX:,1)

("Licensor",1)

(DISCLAIMED.,2)

("Derivative,1)

(elaborations,,1)

(incidental,,1)

(prepare,1)

(A,3)

(exercising,1)

(*/,3)

(which,2)

(pertain,2)

(explicitly,1)

(tort,1)

(3.,1)

(also,1)

(conversions,1)

(liability,2)

(whether,4)

(character,1)

(should,1)

(thereof.,1)

(of,,3)

(your,4)

(royalty-free,,2)

(entities,1)

(or,,1)

(NEGLIGENCE,2)

(author,1)

("Not,1)

(source,9)

(then,2)

((including,3)

(Redistribution.,1)

(attribution,4)

(by,21)

(TO,,4)

(defined,1)

(OWNER,2)

(If,2)

(an,6)

(/*,1)

(Collet.,1)

(improving,1)

(grossly,1)

(COPYRIGHT,4)

(above,,1)

(theory,,1)

(mailing,1)

(7.,1)

(Notwithstanding,1)

(code,,2)

(cross-claim,1)

(provide,1)

((such,1)

(arising,1)

(Object,4)

(In,1)

(-,7)

(those,3)

(work,,2)

(easier,1)

(based,1)

(medium,,1)

(within,8)

(worldwide,,2)

(authorship.,1)

(files.,1)

(inability,1)

(you,2)

(POSSIBILITY,2)

(cannot,1)

(copies,1)

(a,21)

(statement,1)

(above,4)

(state,1)

(work,5)

(by,,3)

(to,41)

(appear.,1)

(Your,9)

(where,1)

(liability.,1)

(governing,1)

(NOT,4)

(License,,6)

(hold,1)

(and,51)

(copyright,15)

(USE,,3)

(compliance,1)

(SOFTWARE,,2)

(comment,1)

(additional,4)

(executed,1)

(mechanical,1)

(Contributor,8)

[[email protected] ~]$

时间: 2024-08-02 00:37:49

Spark API编程动手实战-02-以集群模式进行Spark API实战textFile、cache、count的相关文章

Apache Spark 2.2.0 中文文档 - 集群模式概述 | ApacheCN

集群模式概述 该文档给出了 Spark 如何在集群上运行.使之更容易来理解所涉及到的组件的简短概述.通过阅读 应用提交指南 来学习关于在集群上启动应用. 组件 Spark 应用在集群上作为独立的进程组来运行,在您的 main 程序中通过 SparkContext 来协调(称之为 driver 程序). 具体的说,为了运行在集群上,SparkContext 可以连接至几种类型的 Cluster Manager(既可以用 Spark 自己的 Standlone Cluster Manager,或者

Zookeeper实战之单机集群模式

前一篇文章介绍了Zookeeper的单机模式的安装及应用,但是Zookeeper是为了解决分布式应用场景的,所以通常都会运行在集群模式下.今天由于手头机器不足,所以今天打算在一台机器上部署三个Zookeeper服务来组成一个Zookeeper集群.这里解压Zookeeper的安装包到/opt目录下,这里用三个目录来代表三个Zookeeper实例,分别是/opt/zookeeper1,/opt/zookeeper2和/opt/zookeeper3. 1. 首先编辑每个Zookeeper目录下的co

Spark API编程动手实战-02-以集群模式进行Spark API实战textFile、cach

操作HDFS:先要保证HDFS启动了: 启动spark集群: 以spark-shell运行在spark集群上: 查看下之前上传到HDFS上的"LICENSE.txt"文件: 用spark读取这个文件: 使用count统计该文件的行数: 我们可以看到count 耗时为0.239708s 对该RDD进行cache操作并执行count使得缓存生效: 执行count结果为: 此时耗时为0.21132s 再执行count操作: 此时耗时为0.029580s,这时因为我们自己基于cache后的数据

【原创 Hadoop&Spark 动手实践 5】Spark 基础入门,集群搭建以及Spark Shell

Spark 基础入门,集群搭建以及Spark Shell 主要借助Spark基础的PPT,再加上实际的动手操作来加强概念的理解和实践. Spark 安装部署 理论已经了解的差不多了,接下来是实际动手实验: 练习1 利用Spark Shell(本机模式) 完成WordCount spark-shell 进行Spark-shell本机模式 第一步:通过文件方式导入数据 scala> val rdd1 = sc.textFile("file:///tmp/wordcount.txt")

编写Spark的WordCount程序并提交到集群运行[含scala和java两个版本]

编写Spark的WordCount程序并提交到集群运行[含scala和java两个版本] 1. 开发环境 Jdk 1.7.0_72 Maven 3.2.1 Scala 2.10.6 Spark 1.6.2 Hadoop 2.6.4 IntelliJ IDEA 2016.1.1 2. 创建项目1) 新建Maven项目 2) 在pom文件中导入依赖pom.xml文件内容如下: <?xml version="1.0" encoding="UTF-8"?> &l

第131讲:Hadoop集群管理工具均衡器Balancer 实战详解学习笔记

第131讲:Hadoop集群管理工具均衡器Balancer 实战详解学习笔记 为什么需要均衡器呢? 随着集群运行,具体hdfs各个数据存储节点上的block可能分布得越来越不均衡,会导致运行作业时降低mapreduce的本地性. 分布式计算中精髓性的一名话:数据不动代码动.降低本地性对性能的影响是致使的,而且不能充分利用集群的资源,因为导致任务计算会集中在部分datanode上,更易导致故障. balancer是hadoop的一个守护进程.会将block从忙的datanode移动到闲的datan

Zookeeper实战之嵌入式执行Zookeeper集群模式

非常多使用Zookeeper的情景是须要我们嵌入Zookeeper作为自己的分布式应用系统的一部分来提供分布式服务.此时我们须要通过程序的方式来启动Zookeeper.此时能够通过Zookeeper API的ZooKeeperServerMain类来启动Zookeeper服务. 以下是一个集群模式下启动Zookeeper服务的样例 这里假定我们执行Zookeeper集群的三台机器名分别为fanbinx1,fanbinx2,fanbinx3  首先是zoo.cfg配置文件 tickTime=200

Zookeeper实战之嵌入式运行Zookeeper集群模式

很多使用Zookeeper的情景是需要我们嵌入Zookeeper作为自己的分布式应用系统的一部分来提供分布式服务,此时我们需要通过程序的方式来启动Zookeeper.此时可以通过Zookeeper API的ZooKeeperServerMain类来启动Zookeeper服务. 下面是一个集群模式下启动Zookeeper服务的例子 这里假定我们运行Zookeeper集群的三台机器名分别为fanbinx1,fanbinx2,fanbinx3 首先是zoo.cfg配置文件 tickTime=2000

(7)在集群上运行Spark

7.2 Spark运行时架构 Spark集群采用的是主/从结构.在一个Spark集群中,有一个节点负责中央协调,调度各个分布式工作节点.这个中央协调节点被称为驱动器节点,与之对应的工作节点被称为执行器节点.驱动器节点可以和大量的执行器节点进行通信,他们也都作为独立的Java程序运行. 7.2.1 驱动器节点 Spark驱动器节点是执行你的程序中main方法的进程.其实,当你启动Spark Shell时,你就启动了一个Spark驱动器程序,驱动器程序一旦终止,Spark应用也就结束了.驱动器程序在