作者博客迁移至博客园:http://www.cnblogs.com/xiaodf/
有些应用中可能希望先在driver上运行一段java单机程序,然后再初始化SparkContext用集群模式操作java程序返回值。从而避免过早建立SparkContext对象分配集群资源,使资源长时间空闲。
这里涉及到两个yarn参数:
<property>
<name>yarn.am.liveness-monitor.expiry-interval-ms</name>
<value>6000000</value>
</property>
<property>
<name>yarn.resourcemanager.am.max-retries</name>
<value>10</value>
</property>
Yarn会周期性遍历所有的ApplicationMaster,如果一个ApplicationMaster在一定时间(可通过参数yarn.am.liveness-monitor.expiry-interval-ms配置,默认为10min)内未汇报心跳信息,则认为它死掉了,它上面所有正在运行的Container将被置为运行失败(RM不会重新执行这些Container,它只会通过心跳机制告诉对应的AM,由AM决定是否重新执行,如果需要,则AM重新向RM申请资源),AM本身会被重新分配到另外一个节点上(管理员可通过参数yarn.resourcemanager.am.max-retries指定每个ApplicationMaster的尝试次数,默认是1次)执行。
还需要两个spark参数:
<property>
<name>spark.yarn.am.waitTime</name>
<value>6000000</value>
</property>
<property>
<name>spark.yarn.applicationMaster.waitTries</name>
<value>200</value>
</property>
集群管理
Spark On YARN
属性名称 | 默认值 | 含义 |
---|---|---|
spark.yarn.scheduler.heartbeat.interval-ms | 5000 | Spark AppMaster发送心跳信息给YARN RM的时间间隔 |
spark.yarn.am.waitTime | 100000 | 启动时等待时间 |
spark.yarn.applicationMaster.waitTries | 10 | RM等待Spark AppMaster启动重试次数,也就是SparkContext初始化次数。超过这个数值,启动失败 |
下面是一个测试用例,现在driver打印30分钟的信息,然后再初始化SparkContext
package iie.udps.example.spark;
import iie.udps.common.hcatalog.SerHCatInputFormat;
import iie.udps.common.hcatalog.SerHCatOutputFormat;
import java.io.IOException;
import java.util.HashMap;
import java.util.List;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.WritableComparable;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hive.hcatalog.data.DefaultHCatRecord;
import org.apache.hive.hcatalog.data.HCatRecord;
import org.apache.hive.hcatalog.data.schema.HCatSchema;
import org.apache.hive.hcatalog.mapreduce.OutputJobInfo;
import org.apache.spark.SerializableWritable;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaPairRDD;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.Function;
import org.apache.spark.api.java.function.PairFunction;
import org.apache.spark.api.java.function.Function2;
import scala.Tuple2;
/**
* 实现功能:首先在driver上单机打印30分钟数据,然后初始化SparkContext开启集群模式,用spark+hcatlog 读hive表数据,实现GroupByAge功能,
* 输出结果到hive表中,同时打印xml信息到hdfs文件。
* spark-submit --class iie.udps.example.spark.SparkTest --master yarn-cluster
* --num-executors 2 --executor-memory 1g --executor-cores 1 --driver-memory 1g
* --conf spark.yarn.applicationMaster.waitTries=200,--conf spark.yarn.am.waitTime=1800000 --jars /home/xdf/udps-sdk-0.3.jar,/home/xdf/udps-sdk-0.3.jar
* /home/xdf/sparktest.jar -c /user/hdfs/TestStdin2.xml
*/
public class SparkTest {
@SuppressWarnings("rawtypes")
public static void main(String[] args) throws Exception {
if (args.length < 2) {
System.err.println("Usage: <-c> <stdin.xml>");
System.exit(1);
}
String stdinXml = args[1];
OperatorParamXml operXML = new OperatorParamXml();
List<java.util.Map> stdinList = operXML.parseStdinXml(stdinXml);// 参数列表
// 获得输入参数
String inputDBName = stdinList.get(0).get("inputDBName").toString();
String inputTabName = stdinList.get(0).get("inputTabName").toString();
String outputDBName = stdinList.get(0).get("outputDBName").toString();
String outputTabName = stdinList.get(0).get("outputTabName").toString();
String tempHdfsBasePath = stdinList.get(0).get("tempHdfsBasePath")
.toString();
String jobinstanceid = stdinList.get(0).get("jobinstanceid").toString();
System.out.println(inputDBName+": "+ inputTabName +": "+outputDBName+": "+ outputTabName
+": "+ tempHdfsBasePath+": "+ jobinstanceid);
long begin = System.currentTimeMillis();
int count = 600;// 写文件行数
for (int i = 0; i < count; i++) {
System.out.println("aaaaaaaaaaaaaaa"+i);
Thread.sleep(3000);
}
long end = System.currentTimeMillis();
System.out.println("FileOutputStream执行耗时:" + (end - begin) + "ms");
if (inputDBName == "" || inputTabName == "" || jobinstanceid == ""
|| outputDBName == "" || outputTabName == ""
|| tempHdfsBasePath == "" || jobinstanceid == "") {
// 设置异常输出参数
java.util.Map<String, String> stderrMap = new HashMap<String, String>();
String errorMessage = "Some operating parameters is empty!!!";
String errotCode = "80001";
stderrMap.put("errorMessage", errorMessage);
stderrMap.put("errotCode", errotCode);
stderrMap.put("jobinstanceid", jobinstanceid);
String fileName = "";
if (tempHdfsBasePath.endsWith("/")) {
fileName = tempHdfsBasePath + "stderr.xml";
} else {
fileName = tempHdfsBasePath + "/stderr.xml";
}
// 生成异常输出文件
operXML.genStderrXml(fileName, stderrMap);
} else {
// 根据输入表结构,创建与输入表同样结构的输出表
HCatSchema schema = operXML
.getHCatSchema(inputDBName, inputTabName);
// Spark程序第一件事情就是创建一个JavaSparkContext告诉Spark怎么连接集群
SparkConf sparkConf = new SparkConf().setAppName("SparkExample");
JavaSparkContext jsc = new JavaSparkContext(sparkConf);
// 读取并处理hive表中的数据,生成RDD数据并处理后返回
JavaRDD<SerializableWritable<HCatRecord>> LastRDD = getProcessedData(
jsc, inputDBName, inputTabName, schema);
// 将处理后的数据存到hive输出表中
storeToTable(LastRDD, outputDBName, outputTabName);
jsc.stop();
// 设置正常输出参数
java.util.Map<String, String> stdoutMap = new HashMap<String, String>();
stdoutMap.put("outputDBName", outputDBName);
stdoutMap.put("outputTabName", outputTabName);
stdoutMap.put("jobinstanceid", jobinstanceid);
String fileName = "";
if (tempHdfsBasePath.endsWith("/")) {
fileName = tempHdfsBasePath + "stdout.xml";
} else {
fileName = tempHdfsBasePath + "/stdout.xml";
}
// 生成正常输出文件
operXML.genStdoutXml(fileName, stdoutMap);
}
System.out.println(inputDBName+": "+ inputTabName +": "+outputDBName+": "+ outputTabName
+": "+ tempHdfsBasePath+": "+ jobinstanceid);
System.exit(0);
}
/**
*
* @param jsc
* @param dbName
* @param inputTable
* @param fieldPosition
* @return
* @throws IOException
*/
@SuppressWarnings("rawtypes")
public static JavaRDD<SerializableWritable<HCatRecord>> getProcessedData(
JavaSparkContext jsc, String dbName, String inputTable,
final HCatSchema schema) throws IOException {
// 获取hive表数据
Configuration inputConf = new Configuration();
Job job = Job.getInstance(inputConf);
SerHCatInputFormat.setInput(job.getConfiguration(), dbName, inputTable);
JavaPairRDD<WritableComparable, SerializableWritable> rdd = jsc
.newAPIHadoopRDD(job.getConfiguration(),
SerHCatInputFormat.class, WritableComparable.class,
SerializableWritable.class);
// 获取表记录集
JavaPairRDD<Integer, Integer> pairs = rdd
.mapToPair(new PairFunction<Tuple2<WritableComparable, SerializableWritable>, Integer, Integer>() {
private static final long serialVersionUID = 1L;
@SuppressWarnings("unchecked")
@Override
public Tuple2<Integer, Integer> call(
Tuple2<WritableComparable, SerializableWritable> value)
throws Exception {
HCatRecord record = (HCatRecord) value._2.value();
return new Tuple2((Integer) record.get(1), 1);
}
});
JavaPairRDD<Integer, Integer> counts = pairs
.reduceByKey(new Function2<Integer, Integer, Integer>() {
private static final long serialVersionUID = 1L;
@Override
public Integer call(Integer i1, Integer i2) {
return i1 + i2;
}
});
JavaRDD<SerializableWritable<HCatRecord>> messageRDD = counts
.map(new Function<Tuple2<Integer, Integer>, SerializableWritable<HCatRecord>>() {
private static final long serialVersionUID = 1L;
@Override
public SerializableWritable<HCatRecord> call(
Tuple2<Integer, Integer> arg0) throws Exception {
HCatRecord record = new DefaultHCatRecord(2);
record.set(0, arg0._1);
record.set(1, arg0._2);
return new SerializableWritable<HCatRecord>(record);
}
});
// 返回处理后的数据
return messageRDD;
}
/**
* 将处理后的数据存到输出表中
*
* @param rdd
* @param dbName
* @param tblName
*/
@SuppressWarnings("rawtypes")
public static void storeToTable(
JavaRDD<SerializableWritable<HCatRecord>> rdd, String dbName,
String tblName) {
Job outputJob = null;
try {
outputJob = Job.getInstance();
outputJob.setJobName("SparkExample");
outputJob.setOutputFormatClass(SerHCatOutputFormat.class);
outputJob.setOutputKeyClass(WritableComparable.class);
outputJob.setOutputValueClass(SerializableWritable.class);
SerHCatOutputFormat.setOutput(outputJob,
OutputJobInfo.create(dbName, tblName, null));
HCatSchema schema = SerHCatOutputFormat
.getTableSchemaWithPart(outputJob.getConfiguration());
SerHCatOutputFormat.setSchema(outputJob, schema);
} catch (IOException e) {
e.printStackTrace();
}
// 将RDD存储到目标表中
rdd.mapToPair(
new PairFunction<SerializableWritable<HCatRecord>, WritableComparable, SerializableWritable<HCatRecord>>() {
private static final long serialVersionUID = -4658431554556766962L;
public Tuple2<WritableComparable, SerializableWritable<HCatRecord>> call(
SerializableWritable<HCatRecord> record)
throws Exception {
return new Tuple2<WritableComparable, SerializableWritable<HCatRecord>>(
NullWritable.get(), record);
}
}).saveAsNewAPIHadoopDataset(outputJob.getConfiguration());
}
}
下面是stdin.xml参数文件内容:
<?xml version="1.0" encoding="UTF-8" ?>
<request>
<jobinstanceid>SK9cohJD4yklcD8dJuZXDA</jobinstanceid>
<context>
<property name="userName" value="zhangsan" />
<property name="queueName" value="queue1" />
<property name="processId" value="dns" />
<property name="jobId" value="jobID" />
<property name="hiveServerAddress" value="IP:port " />
<property name="tempDatabaseName" value="database1" />
<property name="tempHdfsBasePath" value="/user/xdf/test/20141216/SK9cohJD4yklcD8dJuZXDA/spark_example_operator" />
<property name="departmentId" value="xx" />
</context>
<operator name="spark_example_operator" alias="sparkExample" class="SK.I.SparkCountByAge">
<parameter name="parse_type">7</parameter>
<parameter name="outputTableName">xdf.test_out</parameter>
</operator>
<datasets>
<dataset name="inport1">
<row>xdf.test_in</row>
</dataset>
</datasets>
</request>
解析xml文件的类如下:
package iie.udps.example.spark;
import iie.udps.common.hcatalog.SerHCatOutputFormat;
import java.io.BufferedReader;
import java.io.File;
import java.io.FileReader;
import java.io.IOException;
import java.io.InputStreamReader;
import java.io.OutputStream;
import java.net.URI;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hive.conf.HiveConf;
import org.apache.hadoop.hive.metastore.HiveMetaStoreClient;
import org.apache.hadoop.hive.metastore.api.FieldSchema;
import org.apache.hadoop.hive.metastore.api.MetaException;
import org.apache.hadoop.hive.metastore.api.SerDeInfo;
import org.apache.hadoop.hive.metastore.api.StorageDescriptor;
import org.apache.hadoop.hive.metastore.api.Table;
import org.apache.hadoop.hive.ql.io.RCFileInputFormat;
import org.apache.hadoop.hive.ql.io.RCFileOutputFormat;
import org.apache.hadoop.hive.serde.serdeConstants;
import org.apache.hadoop.io.WritableComparable;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.util.Progressable;
import org.apache.hive.hcatalog.common.HCatUtil;
import org.apache.hive.hcatalog.data.schema.HCatSchema;
import org.apache.hive.hcatalog.mapreduce.OutputJobInfo;
import org.apache.spark.SerializableWritable;
import org.apache.thrift.TException;
import org.dom4j.Document;
import org.dom4j.DocumentHelper;
import org.dom4j.Element;
import org.dom4j.io.OutputFormat;
import org.dom4j.io.XMLWriter;
/**
* Dom4j 生成XML文档与解析XML文档
*/
public class OperatorParamXml {
@SuppressWarnings("rawtypes")
public List<Map> parseStdinXml(String stdinXml) throws Exception {
// 读取stdin.xml文件
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(conf);
FSDataInputStream dis = fs.open(new Path(stdinXml));
InputStreamReader isr = new InputStreamReader(dis, "utf-8");
BufferedReader read = new BufferedReader(isr);
String tempString = "";
String xmlParams = "";
while ((tempString = read.readLine()) != null) {
xmlParams += "\n" + tempString;
}
read.close();
xmlParams = xmlParams.substring(1);
String userName = null;
String operatorName = null;
String inputDBName = null;
String outputDBName = null;
String inputTabName = null;
String outputTabName = null;
String strs = null;
String fieldName = null;
String inputFilePath = null;
String schemaList = null;
String jobinstanceid = null;
String tempHdfsBasePath = null;
String queueName = null;
String processId = null;
String jobId = null;
String hiveServerAddress = null;
String departmentId = null;
List<Map> list = new ArrayList<Map>();
Map<String, String> map = new HashMap<String, String>();
Document document = DocumentHelper.parseText(xmlParams); // 将字符串转化为xml
Element node1 = document.getRootElement(); // 获得根节点
Iterator iter1 = node1.elementIterator(); // 获取根节点下的子节点
while (iter1.hasNext()) {
Element node2 = (Element) iter1.next();
if ("jobinstanceid".equals(node2.getName())) {
jobinstanceid = node2.getText();
map.put("jobinstanceid", jobinstanceid);
System.out.println("====jobinstanceid=====" + jobinstanceid);
}
// 获取通用参数
if ("context".equals(node2.getName())) {
Iterator iter2 = node2.elementIterator();
while (iter2.hasNext()) {
Element node3 = (Element) iter2.next();
if ("property".equals(node3.getName())) {
if ("userName".equals(node3.attributeValue("name"))) {
userName = node3.attributeValue("value");
map.put("userName", userName);
} else if ("queueName".equals(node3
.attributeValue("name"))) {
queueName = node3.attributeValue("value");
map.put("queueName", queueName);
} else if ("processId".equals(node3
.attributeValue("name"))) {
processId = node3.attributeValue("value");
map.put("processId", processId);
} else if ("jobId".equals(node3.attributeValue("name"))) {
jobId = node3.attributeValue("value");
map.put("jobId", jobId);
} else if ("hiveServerAddress".equals(node3
.attributeValue("name"))) {
hiveServerAddress = node3.attributeValue("value");
map.put("hiveServerAddress", hiveServerAddress);
} else if ("outputDBName".equals(node3
.attributeValue("name"))) {
outputDBName = node3.attributeValue("value");
map.put("outputDBName", outputDBName);
} else if ("tempHdfsBasePath".equals(node3
.attributeValue("name"))) {
tempHdfsBasePath = node3.attributeValue("value");
map.put("tempHdfsBasePath", tempHdfsBasePath);
} else if ("departmentId".equals(node3
.attributeValue("name"))) {
departmentId = node3.attributeValue("value");
map.put("departmentId", departmentId);
}
}
}
}
// 获取算子参数
if ("operator".equals(node2.getName())) {
operatorName = node2.attributeValue("name");
map.put("operatorName", operatorName);
Iterator iter2 = node2.elementIterator();
while (iter2.hasNext()) {
Element node3 = (Element) iter2.next();
if ("parameter".equals(node3.getName())) {
if ("field1".equals(node3.attributeValue("name"))) {
fieldName = node3.getText();
map.put("fieldName", fieldName);
}
if ("outputTableName".equals(node3.attributeValue("name"))) {
String tempStr = node3.getText(); // 获得数据库.表格式字符串
if (!"".equals(tempStr.trim())) {
String[] arr = tempStr.split("\\.");
outputDBName = arr[0];
outputTabName = arr[1];
}
map.put("outputDBName", outputDBName);
map.put("outputTabName", outputTabName);
}
if ("inputFilePath"
.equals(node3.attributeValue("name"))) {
inputFilePath = node3.getText();
map.put("inputFilePath", inputFilePath);
}
if ("schemaList".equals(node3.attributeValue("name"))) {
schemaList = node3.getText();
map.put("schemaList", schemaList);
}
}
}
}
// 获取输入数据库
if ("datasets".equals(node2.getName())) {
Iterator iter2 = node2.elementIterator();
while (iter2.hasNext()) {
Element node3 = (Element) iter2.next();
if ("inport1".equals(node3.attributeValue("name"))) {
Iterator iter3 = node3.elementIterator();
while (iter3.hasNext()) {
Element node4 = (Element) iter3.next();
strs = node4.getText();
}
if (!"".equals(strs.trim())) {
String[] arr = strs.split("\\.");
if(arr.length == 2){
inputDBName = arr[0];
inputTabName = arr[1];
}
}
map.put("inputDBName", inputDBName);
map.put("inputTabName", inputTabName);
}
}
}
}
list.add(map);
return list;
}
@SuppressWarnings("rawtypes")
public List<Map> parseStdoutXml(String stdinXml) throws Exception {
// 读取stdin.xml文件
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(conf);
FSDataInputStream dis = fs.open(new Path(stdinXml));
InputStreamReader isr = new InputStreamReader(dis, "utf-8");
BufferedReader read = new BufferedReader(isr);
String tempString = "";
String xmlParams = "";
while ((tempString = read.readLine()) != null) {
xmlParams += "\n" + tempString;
}
read.close();
xmlParams = xmlParams.substring(1);
String outputDBName = null;
String outputTabName = null;
String jobinstanceid = null;
List<Map> list = new ArrayList<Map>();
Map<String, String> map = new HashMap<String, String>();
Document document = DocumentHelper.parseText(xmlParams); // 将字符串转化为xml
Element node1 = document.getRootElement(); // 获得根节点
Iterator iter1 = node1.elementIterator(); // 获取根节点下的子节点
while (iter1.hasNext()) {
Element node2 = (Element) iter1.next();
if ("jobinstanceid".equals(node2.getName())) {
jobinstanceid = node2.getText();
map.put("jobinstanceid", jobinstanceid);
System.out.println("====jobinstanceid=====" + jobinstanceid);
}
// 获取输入数据库
if ("datasets".equals(node2.getName())) {
Iterator iter2 = node2.elementIterator();
while (iter2.hasNext()) {
Element node3 = (Element) iter2.next();
if ("outport1".equals(node3.attributeValue("name"))) {
Iterator iter3 = node3.elementIterator();
String strs = null;
while (iter3.hasNext()) {
Element node4 = (Element) iter3.next();
strs = node4.getText();
}
if (!"".equals(strs.trim())) {
String[] arr = strs.split("\\.");
outputDBName = arr[0];
outputTabName = arr[1];
}
map.put("outputDBName", outputDBName);
map.put("outputTabName", outputTabName);
}
}
}
}
list.add(map);
return list;
}
/* 生成stdout.Xml文件 */
@SuppressWarnings("rawtypes")
public void genStdoutXml(String fileName, Map listOut) {
String jobinstance = null;
String outputDBName = null;
String outputTabName = null;
jobinstance = listOut.get("jobinstanceid").toString();
outputDBName = listOut.get("outputDBName").toString();
outputTabName = listOut.get("outputTabName").toString();
Document document = DocumentHelper.createDocument();
Element response = document.addElement("response");
Element jobinstanceid = response.addElement("jobinstanceid");
jobinstanceid.setText(jobinstance);
Element datasets = response.addElement("datasets");
Element dataset = datasets.addElement("dataset");
dataset.addAttribute("name", "outport1");
Element row = dataset.addElement("row");
row.setText(outputDBName + "." + outputTabName);
try {
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(URI.create(fileName), conf);
OutputStream out = fs.create(new Path(fileName),
new Progressable() {
public void progress() {
}
});
OutputFormat format = OutputFormat.createPrettyPrint();
format.setEncoding("UTF-8");
XMLWriter xmlWriter = new XMLWriter(out, format);
xmlWriter.write(document);
xmlWriter.close();
} catch (IOException e) {
System.out.println(e.getMessage());
}
}
/* 生成stderr.xml文件 */
@SuppressWarnings("rawtypes")
public void genStderrXml(String fileName, Map listOut) {
String jobinstance = null;
String errorMessage = null;
String errotCode = null;
jobinstance = listOut.get("jobinstanceid").toString();
errorMessage = listOut.get("errorMessage").toString();
errotCode = listOut.get("errotCode").toString();
Document document = DocumentHelper.createDocument();
Element response = document.addElement("error");
Element jobinstanceid = response.addElement("jobinstanceid");
jobinstanceid.setText(jobinstance);
Element code = response.addElement("code");
code.setText(errotCode);
Element message = response.addElement("message");
message.setText(errorMessage);
try {
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(URI.create(fileName), conf);
OutputStream out = fs.create(new Path(fileName),
new Progressable() {
public void progress() {
}
});
OutputFormat format = OutputFormat.createPrettyPrint();
format.setEncoding("UTF-8");
XMLWriter xmlWriter = new XMLWriter(out, format);
xmlWriter.write(document);
xmlWriter.close();
} catch (IOException e) {
System.out.println(e.getMessage());
}
}
/**
* 获得表模式
*
* @param dbName
* @param tblName
* @return
*/
public HCatSchema getHCatSchema(String dbName, String tblName) {
Job outputJob = null;
HCatSchema schema = null;
try {
outputJob = Job.getInstance();
outputJob.setJobName("getHCatSchema");
outputJob.setOutputFormatClass(SerHCatOutputFormat.class);
outputJob.setOutputKeyClass(WritableComparable.class);
outputJob.setOutputValueClass(SerializableWritable.class);
SerHCatOutputFormat.setOutput(outputJob,
OutputJobInfo.create(dbName, tblName, null));
schema = SerHCatOutputFormat.getTableSchema(outputJob
.getConfiguration());
} catch (IOException e) {
e.printStackTrace();
}
return schema;
}
/**
* 指定数据库、表名,创建表
*
* @param dbName
* @param tblName
* @param schema
*/
public void createTable(String dbName, String tblName, HCatSchema schema) {
HiveMetaStoreClient client = null;
try {
HiveConf hiveConf = HCatUtil.getHiveConf(new Configuration());
client = HCatUtil.getHiveClient(hiveConf);
} catch (MetaException | IOException e) {
e.printStackTrace();
}
try {
if (client.tableExists(dbName, tblName)) {
client.dropTable(dbName, tblName);
}
} catch (TException e) {
e.printStackTrace();
}
// 获取表模式
List<FieldSchema> fields = HCatUtil.getFieldSchemaList(schema
.getFields());
// 生成表对象
Table table = new Table();
table.setDbName(dbName);
table.setTableName(tblName);
StorageDescriptor sd = new StorageDescriptor();
sd.setCols(fields);
table.setSd(sd);
sd.setInputFormat(RCFileInputFormat.class.getName());
sd.setOutputFormat(RCFileOutputFormat.class.getName());
sd.setParameters(new HashMap<String, String>());
sd.setSerdeInfo(new SerDeInfo());
sd.getSerdeInfo().setName(table.getTableName());
sd.getSerdeInfo().setParameters(new HashMap<String, String>());
sd.getSerdeInfo().getParameters()
.put(serdeConstants.SERIALIZATION_FORMAT, "1");
sd.getSerdeInfo().setSerializationLib(
org.apache.hadoop.hive.serde2.columnar.ColumnarSerDe.class
.getName());
Map<String, String> tableParams = new HashMap<String, String>();
table.setParameters(tableParams);
try {
client.createTable(table);
System.out.println("Create table successfully!");
} catch (TException e) {
e.printStackTrace();
return;
} finally {
client.close();
}
}
/**
* 以行为单位读取文件,常用于读面向行的格式化文件
*/
public static void readFileByLines(String fileName) {
File file = new File(fileName);
BufferedReader reader = null;
try {
System.out.println("以行为单位读取文件内容,一次读一整行:");
reader = new BufferedReader(new FileReader(file));
String tempString = null;
int line = 1;
// 一次读入一行,直到读入null为文件结束
while ((tempString = reader.readLine()) != null) {
// 显示行号
System.out.println("line " + line + ": " + tempString);
line++;
}
reader.close();
} catch (IOException e) {
e.printStackTrace();
} finally {
if (reader != null) {
try {
reader.close();
} catch (IOException e1) {
}
}
}
}
}
输入表数据:
hive> select * from test_in;
OK
1 20
2 20
3 21
4 20
5 21
6 20
7 21
8 19
9 19
10 21
输出表数据:
hive> select * from test_out;
OK
19 2
21 4
20 4