Sqoop1.99.4 Java API实践

标签：

软件版本：

Sqoop：1.99.4；JDK：1.7；Hadoop2.2；

环境：

1. Hadoop集群：

node12:NameNode、ResourceManager、JobHistoryServer、DataNode、NodeManager； 1.5G内存+10G硬盘+1核

node13:SecondaryNameNode、DataNode、NodeManager；1.5G内存+10G硬盘+1核

（1）找到sqoop-1.99.4-bin-hadoop200\server\webapps\sqoop.war，接着，解压sqoop.war文件，找到：

sqoop-1.99.4-bin-hadoop200\server\webapps\sqoop\WEB-INF\lib目录，拷贝下面的lib包到工程build path（非运行必须）；

（2）找到目录sqoop-1.99.4-bin-hadoop200\shell\lib ，拷贝下面的lib包到工程的build path（运行必须）

Sqoop简单概念：

1. Sqoop connector：

Sqoop连接器，sqoop server启动后，使用客户端输入命令：show connector，可以看到sqoop目前的连接器，如下所示：

2. Sqoop link：

一个link需要一个connector的支持，所以link的建立需要指定一个connector（指定其id即可，从下面的代码可以看到）；link是对connector的一层封装可以设置参数，比如针对mysql，可以设置mysql driver的名字、链接字符串、用户名、密码等等参数；针对hdfs，可以设置uri；具体可以设置的参数如下所示：

hdfs:

Link configuration:
linkConfig.uri,HDFS URI : null

From Job configuration:
fromJobConfig.inputDirectory,Input directory : null

ToJob configuration:
toJobConfig.outputFormat,Output format : null
toJobConfig.compression,Compression format : null
toJobConfig.customCompression,Custom compression format : null
toJobConfig.outputDirectory,Output directory : null

Link configuration:
linkConfig.jdbcDriver,JDBC Driver Class : null
linkConfig.connectionString,JDBC Connection String : null
linkConfig.username,Username : null
linkConfig.password,Password : null
linkConfig.jdbcProperties,JDBC Connection Properties : null

From database configuration:
fromJobConfig.schemaName,Schema name : null
fromJobConfig.tableName,Table name : null
fromJobConfig.sql,Table SQL statement : null
fromJobConfig.columns,Table column names : null
fromJobConfig.partitionColumn,Partition column name : null
fromJobConfig.allowNullValueInPartitionColumn,Null value allowed for the partition column : null
fromJobConfig.boundaryQuery,Boundary query : null

To database configuration:
toJobConfig.schemaName,Schema name : null
toJobConfig.tableName,Table name : null
toJobConfig.sql,Table SQL statement : null
toJobConfig.columns,Table column names : null
toJobConfig.stageTableName,Stage table name : null
toJobConfig.shouldClearStageTable,Should clear stage table : null

3. Sqoop job：

Sqoop任务的建立需要指定两个东西，其一：指定from和to，from和to只需要提供其对应的link的id即可；其二，则是指定from和to的具体参数（这里的参数是指变动比较大的参数，比如hdfs中的输入目录，jdbc中的表名或列字符串等）；

Sqoop Java API：

0. 建立connector（这个不需要建立，在配置好sqoop后自然会有）

1. 建立link

1）在建立link前，先在sqoop client中查看现有的link：

show link ，可以看到是没有link的；

2）建立hdfs的link：

//        long connectorId = 1;
//        MLink link = createHDFSLink(client,connectorId);

 long connectorId=2;
//        MLink link = createMysqlLink(client,connectorId);

2. 建立任务：

0）查看现有的任务：show job ，可以看到是没有的（默认没有）

1）建立job：

// 建立任务
//        long fromLinkId=1;
//        long toLinkId=2;
//        String input="/user/fansy/name.data";
//        String tableName="test_name_age";
//        String columns="id,name,age";
//        int mappers = 2;
//        MJob job = createJob(client,fromLinkId,toLinkId,input,tableName,columns,mappers);
//        System.out.println(job.getEnabled());

3. 启动任务：

// 启动任务
//        long jobId =1;
//        startJob(client,jobId);

所有代码如下所示：

package sqoop;

import org.apache.sqoop.client.SqoopClient;
import org.apache.sqoop.model.*;
import org.apache.sqoop.submission.counter.Counter;
import org.apache.sqoop.submission.counter.CounterGroup;
import org.apache.sqoop.submission.counter.Counters;
import org.apache.sqoop.validation.Status;

import java.util.List;
import java.util.ResourceBundle;

/**
 * Created by fansy on 2016/3/7.
 */
public class Sqoop1_99_4_Test {

    public static void main(String[] args){
        String url = "http://node13:12000/sqoop/";
        SqoopClient client = new SqoopClient(url);

        // 建立link 连接
//        long connectorId = 1;
//        MLink link = createHDFSLink(client,connectorId);
        long connectorId=2;
//        MLink link = createMysqlLink(client,connectorId);
        listInputAndOutputParameters(client,connectorId);
// 建立任务
//        long fromLinkId=1;
//        long toLinkId=2;
//        String input="/user/fansy/name.data";
//        String tableName="test_name_age";
//        String columns="name,age";
//        int mappers = 2;
//        MJob job = createJob(client,fromLinkId,toLinkId,input,tableName,columns,mappers);
//        System.out.println(job.getEnabled());
// 启动任务
//        long jobId =1;
//        startJob(client,jobId);
    }

    static void describe(List<MConfig> configs, ResourceBundle resource) {
        for (MConfig config : configs) {
            System.out.println(resource.getString(config.getLabelKey()) + ":");
            List<MInput<?>> inputs = config.getInputs();
            for (MInput input : inputs) {
                System.out.println(input.getName()+","+resource.getString(input.getLabelKey()) + " : " + input.getValue());
            }
            System.out.println();
        }
    }

    /**
     * 打印各个connector的参数
     * @param client
     * @param connectorId
     */
    static void listInputAndOutputParameters(SqoopClient client,long connectorId){

//        String url = "http://node13:12000/sqoop/";
//        SqoopClient client = new SqoopClient(url);

//        long connectorId = link.getConnectorId();

// link config for connector
        describe(client.getConnector(connectorId).getLinkConfig().getConfigs(), client.getConnectorConfigBundle(connectorId));
// from job config for connector
        describe(client.getConnector(connectorId).getFromConfig().getConfigs(), client.getConnectorConfigBundle(connectorId));
// to job config for the connector
        describe(client.getConnector(connectorId).getToConfig().getConfigs(), client.getConnectorConfigBundle(connectorId));
        // create a placeholder for link
    }

    /**
     * 建立mysql link
     * @param client
     * @param connectorId
     * @return
     */
    static MLink createMysqlLink(SqoopClient client,long connectorId){
        MLink link = client.createLink(connectorId);
        link.setName("mysql");
        link.setCreationUser("fansy");
        MLinkConfig linkConfig = link.getConnectorLinkConfig();
// fill in the link config values
        linkConfig.getStringInput("linkConfig.connectionString").setValue("jdbc:mysql://192.168.0.1/test_fz");
        linkConfig.getStringInput("linkConfig.jdbcDriver").setValue("com.mysql.jdbc.Driver");
        linkConfig.getStringInput("linkConfig.username").setValue("fansy");
        linkConfig.getStringInput("linkConfig.password").setValue("fansy");
// save the link object that was filled
        Status status = client.saveLink(link);
        if(status.canProceed()) {
            System.out.println("Created Link with Link Id : " + link.getPersistenceId());
        } else {
            System.out.println("Something went wrong creating the link");
        }
        return link;
    }

    /**
     * 建立HDFS link
     * @param client
     * @param connectorId
     * @return
     */
    static MLink createHDFSLink(SqoopClient client,long connectorId){
        MLink link = client.createLink(connectorId);
        link.setName("hdfs");
        link.setCreationUser("fansy");
        MLinkConfig linkConfig = link.getConnectorLinkConfig();
// fill in the link config values
        linkConfig.getStringInput("linkConfig.uri").setValue("hdfs://node12:8020");
// save the link object that was filled
        Status status = client.saveLink(link);
        if(status.canProceed()) {
            System.out.println("Created Link with Link Id : " + link.getPersistenceId());
        } else {
            System.out.println("Something went wrong creating the link");
        }
        return link;
    }

    /**
     * 建立hdfs to mysql 任务
     * @param client
     * @param fromLinkId
     * @param toLinkId
     * @param input
     * @param tableName
     * @param columns
     * @param mappers
     * @return
     */
    static MJob createJob(SqoopClient client,long fromLinkId,long toLinkId,
                          String input,String tableName,String columns,int mappers){
        MJob job = client.createJob(fromLinkId, toLinkId);
        job.setName("hdfs to mysql");
        job.setCreationUser("fansy");
// set the "FROM" link job config values
        MFromConfig fromJobConfig = job.getFromJobConfig();
        fromJobConfig.getStringInput("fromJobConfig.inputDirectory").setValue(input);

// set the "TO" link job config values
        MToConfig toJobConfig = job.getToJobConfig();
        toJobConfig.getStringInput("toJobConfig.tableName").setValue(tableName);
        toJobConfig.getStringInput("toJobConfig.columns").setValue(columns);

// set the driver config values
        MDriverConfig driverConfig = job.getDriverConfig();
//        driverConfig.getStringInput("throttlingConfig.numExtractors").setValue(String.valueOf(mappers));
        driverConfig.getIntegerInput("throttlingConfig.numExtractors").setValue(mappers);
        Status status = client.saveJob(job);
        if(status.canProceed()) {
            System.out.println("Created Job with Job Id: "+ job.getPersistenceId());
        } else {
            System.out.println("Something went wrong creating the job");
        }
        return job;
    }

    /**
     *  启动任务
     * @param client
     * @param jobId
     */
    static void startJob(SqoopClient client,long jobId){
        MSubmission submission = client.startJob(jobId);
        System.out.println("Job Submission Status : " + submission.getStatus());
        if(submission.getStatus().isRunning() && submission.getProgress() != -1) {
            System.out.println("Progress : " + String.format("%.2f %%", submission.getProgress() * 100));
        }
        System.out.println("Hadoop job id :" + submission.getExternalId());
        System.out.println("Job link : " + submission.getExternalLink());
        Counters counters = submission.getCounters();
        if(counters != null) {
            System.out.println("Counters:");
            for(CounterGroup group : counters) {
                System.out.print("\t");
                System.out.println(group.getName());
                for(Counter counter : group) {
                    System.out.print("\t\t");
                    System.out.print(counter.getName());
                    System.out.print(": ");
                    System.out.println(counter.getValue());
                }
            }
        }
        if(submission.getExceptionInfo() != null) {
            System.out.println("Exception info : " +submission.getExceptionInfo());
        }



    }

    /**
     * 查看任务状态
     * @param client
     * @param jobId
     */
    static void checkJobStatus(SqoopClient client,long jobId){
        //Check job status for a running job
        MSubmission submission = client.getJobStatus(jobId);
        if(submission.getStatus().isRunning() && submission.getProgress() != -1) {
            System.out.println("Progress : " + String.format("%.2f %%", submission.getProgress() * 100));
        }
        System.out.println("Job status:"+submission.getStatus().name());
    }


}

1. HDFS数据：

11kate,3
fansy,22
12kate,3
tom,32
1kate,3
2kate,3

CREATE TABLE `test_name_age` (
  `name` varchar(255) DEFAULT NULL,
  `age` int(11) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

总结：

1. 数据缺失：

可以看到数据库中的数据name列数据前后都少了一个字符，但是在8088端口看到的日志中看到读取的数据确实是完整的：

2. 不能指定分隔符，Sqoop1.99.4版本只有一个默认的分隔符实现

CSVIntermediateDataFormat

如果要用其他的分隔符，那么就要自己去实现；在sqoop1.99.6版本（目前最新版）中支持了json和avro的实现，如下：

最终一句话，sqoop1.99的版本还是bug多多，实际中还是用回1.4的版本吧！

Sqoop1.99.4 Java API实践

标签：

原文地址：http://blog.csdn.net/fansy1990/article/details/50826543