简介
ETL是英文Extract-Transform-Load的缩写,用来描述将数据从源端经过抽取(extract)、转换(transform)、加载(load)至目的端的过程,它能够对各种分布的、异构的源数据(如关系数据)进行抽取,按照预先设计的规则将不完整数据、重复数据以及错误数据等“脏"数据内容进行清洗,得到符合要求的“干净”数据,并加载到数据仓库中进行存储,这些“干净”数据就成为了数据分析、数据挖掘的基石。
kettle是一个开源ETL工具。kettle提供了基于java的图形化界面,使用很方便。kettle提供了基于 JAVA的脚步编写功能,可以灵活地自定义ETL过程,使自行定制、批量处理等成为可能,这才是一个程序员需要做的工作,而不仅是象使用word一样操作 kettle用户界面。
环境集成:
参考:java集成kettle教程(附示例代码)_kettle java_成伟平2022的博客-CSDN博客
代码:
pom.xml添加:
<!--mysql数据库链接驱动以及连接池-->
<dependency>
<groupId>mysql</groupId>
<artifactId>mysql-connector-java</artifactId>
</dependency>
<dependency>
<groupId>com.alibaba</groupId>
<artifactId>druid</artifactId>
<version>1.2.11</version>
</dependency>
<!-- kettle 工具本地jar包加载 -->
<dependency>
<groupId>pentaho-kettle</groupId>
<artifactId>kettle-core</artifactId>
<version>8.2.0.7-719</version>
<scope>system</scope>
<systemPath>${project.basedir}/lib/kettle-core-8.2.0.7-719.jar</systemPath>
</dependency>
<dependency>
<groupId>pentaho-kettle</groupId>
<artifactId>kettle-engine</artifactId>
<version>8.2.0.7-719</version>
<scope>system</scope>
<systemPath>${project.basedir}/lib/kettle-engine-8.2.0.7-719.jar</systemPath>
</dependency>
<dependency>
<groupId>pentaho-kettle</groupId>
<artifactId>metastore</artifactId>
<version>8.2.0.7-719</version>
<scope>system</scope>
<systemPath>${project.basedir}/lib/metastore-8.2.0.7-719.jar</systemPath>
</dependency>
<!--kettle需要用到的其它依赖-->
<dependency>
<groupId>org.apache.commons</groupId>
<artifactId>commons-vfs2</artifactId>
<version>2.2</version>
</dependency>
<dependency>
<groupId>com.google.guava</groupId>
<artifactId>guava</artifactId>
<version>17.0</version>
</dependency>
<dependency>
<groupId>commons-io</groupId>
<artifactId>commons-io</artifactId>
<version>2.2</version>
</dependency>
<dependency>
<groupId>commons-lang</groupId>
<artifactId>commons-lang</artifactId>
<version>2.6</version>
</dependency>
<dependency>
<groupId>commons-codec</groupId>
<artifactId>commons-codec</artifactId>
<version>1.10</version>
</dependency>
<dependency>
<groupId>com.jcraft</groupId>
<artifactId>jsch</artifactId>
<version>0.1.54</version>
</dependency>
<dependency>
<groupId>net.sourceforge.jexcelapi</groupId>
<artifactId>jxl</artifactId>
<version>2.6.12</version>
</dependency>
@RestController
@RequestMapping("${application.admin-path}/etl-kettl")
//@Api(tags = "ETL-Kettle的demo接口")
public class KettleDemoContrllor {
@Resource
KettleService kettleService;
@GetMapping("/execKtr")
//@ApiOperation("执行ktr文件")
private Object runKtr(String filename) throws Exception {
return R.buildOkData(kettleService.runTaskKtr(filename,null).toString());
}
@GetMapping("/execKjb")
//@ApiOperation("执行kjb文件")
private Object runKjb(String filename) throws Exception {
return R.buildOkData(kettleService.runTaskKjb(filename, null).toString());
}
}
public interface KettleService {
/**
* 开始执行ETL任务(ktr文件)
*
* @param taskFileName 执行的任务文件名(ktr)
* @param params 执行任务输入的参数
* @return 运行结果
* @throws Exception 没有找到配置文件,Kettle的运行异常不会抛出
*/
Object runTaskKtr(String taskFileName, Map<String, String> params) throws Exception;
/**
* 开始执行ETL任务(kjb文件)
*
* @param taskFileName 执行的任务文件名(kjb)
* @param params 执行任务输入的参数
* @return 运行结果
* @throws Exception 没有找到配置文件,Kettle的运行异常不会抛出
*/
Object runTaskKjb(String taskFileName, Map<String, String> params) throws Exception;
}
@Service
public class KettleServiceImpl implements KettleService {
@Value("${kettle.script.path}")
private String kettleScriptPath;
private static final Logger logger = LoggerFactory.getLogger("kettle-service-log");
private final List<KtrMeta> KTR_METAS = new ArrayList<>();
private final List<KjbMeta> KJB_METAS = new ArrayList<>();
private List<String> getFiles(String path, String subName) {
List<String> files = new ArrayList<>();
File file = new File(path);
File[] tempList = file.listFiles();
if (tempList == null){
return files;
}
for (File value : tempList) {
if (value.isFile()) {
if (Objects.equals(value.toString().substring(value.toString().length() - 3), subName)) {
files.add(value.getName());
}
}
}
return files;
}
//采用单列模式,项目启动时加载环境,加载所有的转换配置、任务配置,后续执行就会快一点
//@PostConstruct
public void init() throws KettleException {
logger.info("----------------------开始初始化ETL配置------------------------");
KettleEnvironment.init();
List<String> ktrFiles = getFiles(kettleScriptPath, "ktr");
List<String> kjbFiles = getFiles(kettleScriptPath, "kjb");
logger.info("需要加载的转换为:" + ktrFiles.toString());
logger.info("需要加载的任务为:" + kjbFiles.toString());
logger.info("----------------------开始加载ETL配置--------------------------");
for (String ktrFile : ktrFiles) {
KtrMeta ktrMeta = new KtrMeta();
ktrMeta.setName(ktrFile);
ktrMeta.setTransMeta(new TransMeta(kettleScriptPath + ktrFile));
KTR_METAS.add(ktrMeta);
logger.info("成功加载转换配置:" + ktrFile);
}
for (String kjbFile : kjbFiles) {
KjbMeta kjbMeta = new KjbMeta();
kjbMeta.setName(kjbFile);
kjbMeta.setJobMeta(new JobMeta(kettleScriptPath + kjbFile, null));
KJB_METAS.add(kjbMeta);
logger.info("成功加载任务配置:" + kjbFile);
}
logger.info("----------------------全部ETL配置加载完毕-----------------------");
}
@Override
public Object runTaskKtr(String ktrFileName, Map<String, String> params) {
logger.info("开始执行转换:" + ktrFileName);
TransMeta transMeta = null;
for (KtrMeta ktrMeta : KTR_METAS) {
if(Objects.equals(ktrFileName,ktrMeta.getName())){
transMeta = ktrMeta.getTransMeta();
break;
}
}
//如果在缓存的列表里面没找到需要自信的配置,尝试手动加载
try {
if (transMeta == null) {
logger.warn("资源池没有找到配置文件:" + ktrFileName+" 尝试二次加载!");
KettleEnvironment.init();
transMeta = new TransMeta(kettleScriptPath + File.separator + ktrFileName);
if(transMeta==null) throw new RuntimeException("未找到需要执行的转换配置文件:");
}
Trans trans = new Trans(transMeta);
if (params != null) {
for (Map.Entry<String, String> entry : params.entrySet()) {
trans.setParameterValue(entry.getKey(), entry.getValue());
}
}
//trans.prepareExecution(null);
//trans.startThreads(); //启用新的线程加载
trans.execute(null);
trans.waitUntilFinished();
return trans.getResult();
}catch (Exception e)
{
e.printStackTrace();
return e.getMessage();
}
}
@Override
public Object runTaskKjb(String objFileName, Map<String, String> params) throws Exception {
logger.info("开始执行任务:" + objFileName);
JobMeta jobMeta = null;
for (KjbMeta kjbMeta : KJB_METAS) {
if(Objects.equals(objFileName,kjbMeta.getName())){
jobMeta = kjbMeta.getJobMeta();
}
}
try {
if (jobMeta == null) {
logger.warn("资源池没有找到配置文件:" + objFileName+" 尝试二次加载!");
KettleEnvironment.init();
jobMeta = new JobMeta(kettleScriptPath + File.separator + objFileName,null);
if(jobMeta==null) throw new RuntimeException("未找到需要执行的任务配置文件:"+objFileName);
}
Job job = new Job(null, jobMeta);
if (params != null) {
for (Map.Entry<String, String> entry : params.entrySet()) {
job.setParameterValue(entry.getKey(), entry.getValue());
}
}
job.start();
job.waitUntilFinished();
return job.getResult();
}catch (Exception e)
{
e.printStackTrace();
return e.getMessage();
}
}
}
@Data
public class KtrMeta {
private TransMeta transMeta;
private String name;
}
@Data
public class KjbMeta {
private JobMeta jobMeta;
private String name;
}
总结:
集成后感觉没什么必要集成到项目里面去。关键还是需要学会工具的使用,以便进行数据收集与治理。
参考:1_ETL和Kettle概述_哔哩哔哩_bilibili
下载: kettle工具下载
版权归原作者 萤火1129 所有, 如有侵权,请联系我们删除。