sparkSQL以JDBC为数据源

一、环境准备

安装oracle后，创建测试表、数据：

1. create table test (
2.  username varchar2(32) primary key ,
3.  password varchar2(32)
4. ); 
5. insert into test values('John','1234');
6. insert into test values('Mike','1234');
7. insert into test values('Jim','1234');
8. insert into test values('Ana','1234');
9. insert into test values('Ennerson','1234');
10. commit;

 

二、实现代码

1、建立JDBC连接读取数据

1. SparkConf sparkConf = new SparkConf().setAppName("JavaSparkSQL").setMaster("local[6]");
2. JavaSparkContext jsc = new JavaSparkContext(sparkConf);
3. SQLContext sqlContext = new SQLContext(jsc);
4.  
5. Map<String, String> options = new HashMap<String, String>();
6. options.put("url", "jdbc:oracle:thin:@192.168.168.100:1521/orcl");
7. options.put("user", "flume");
8. options.put("password","1234");
9. //读取test表
10. options.put("dbtable", "test");
11. Dataset<Row> df = sqlContext.read().format("jdbc").options(options).load();
12. df.show();
13. /*+--------+--------+
14. |USERNAME|PASSWORD|
15. +--------+--------+
16. |    John|    1234|
17. |    Mike|    1234|
18. |     Jim|    1234|
19. |     Ana|    1234|
20. |Ennerson|    1234|
21. +--------+--------+*/

2、遍历Dataset<Row>集合

1. //遍历Dataset<Row>集合
2. List<Row> list = df.collectAsList();
3. //读取test表中username字段的数据
4. for(int i = 0;i < list.size();i++){
5. System.out.println(list.get(i).<String>getAs("USERNAME"));
6. }
7. /*John
8. Mike
9. Jim
10. Ana
11. Ennerson*/

3、执行SQL语句

1. //执行sql语句
2. //一定要有df.createOrReplaceTempView("test"); 否则会报
3. //“Exception in thread "main" org.apache.spark.sql.AnalysisException: Table or view not found: test; line 1 pos 0”
4. df.createOrReplaceTempView("test");
5. sqlContext.sql("insert into test values('Obama','6666')");

4、引入spark-sql依赖包

在pom.xml文件中引入sparksql依赖包

1. <!-- https://mvnrepository.com/artifact/org.apache.spark/spark-sql_2.11 -->
2. <dependency>
3.     <groupId>org.apache.spark</groupId>
4.     <artifactId>spark-sql_2.11</artifactId>
5.     <version>2.1.1</version>
6.     <scope>runtime</scope>
7. </dependency>

三、No suitable driver解决

1、在Eclipse上运行报Exception in thread "main" java.sql.SQLException: No suitable driver错误：

1. Exception in thread "main" java.sql.SQLException: No suitable driver    at java.sql    .DriverManager    .getDriver(DriverManager.java:315)    at org.apache.spark.sql.execution.datasources.jdbc    .JDBCOptions$$anonfun$7    .apply(JDBCOptions.scala:84)    at org.apache.spark.sql.execution.datasources.jdbc    .JDBCOptions$$anonfun$7.apply    (JDBCOptions.scala:84)    at scala.Option.getOrElse(Option.scala:121    )    at org.apache.spark.sql.execution    .datasources.jdbc.JDBCOptions.<init>(JDBCOptions.scala:83)    at org    .apache.spark.sql.execution.datasources    .jdbc.JDBCOptions.<init>(JDBCOptions.scala:34)    at org.apache    .spark.sql.execution.datasources.jdbc    .JdbcRelationProvider.createRelation(JdbcRelationProvider            .scala:32)    at org.apache.spark.sql    .execution.datasources.DataSource.resolveRelation(DataSource.scala:330    )    at org.apache.spark.sql    .DataFrameReader.load(DataFrameReader.scala:152)    at org.apache.spark.sql    .DataFrameReader.load    (DataFrameReader.scala:125)    at com.spark.test.JavaSparkSQL.main(JavaSparkSQL.java    :26    )

原因是没有引用oracle的jdbc驱动包，配置pom.xml文件如下：

1. <!-- oracle jdbc驱动 -->
2. <dependency>      
3.   <groupId>com.oracle</groupId>      
4.   <artifactId>ojdbc5</artifactId>      
5.   <version>11.2.0.1.0</version>
6.   <scope>runtime</scope>
7. </dependency>  

由于Oracle授权问题，Maven不提供oracle JDBC driver，为了在Maven项目中应用Oracle JDBC driver,必须手动添加到本地仓库。

具体可以参考：maven添加oracle jdbc依赖

2、在spark集群环境上运行报Exception in thread "main" java.sql.SQLException: No suitable driver错误：

1. Exception in thread "main" java.sql.SQLException: No suitable driver
2. at java.sql.DriverManager.getDriver(DriverManager.java:315)
3. at org.apache.spark.sql.execution.datasources.jdbc.JDBCOptions$$anonfun$7.apply(JDBCOptions.scala:84)
4. at org.apache.spark.sql.execution.datasources.jdbc.JDBCOptions$$anonfun$7.apply(JDBCOptions.scala:84)
5. at scala.Option.getOrElse(Option.scala:121)
6. at org.apache.spark.sql.execution.datasources.jdbc.JDBCOptions.<init>(JDBCOptions.scala:83)
7. at org.apache.spark.sql.execution.datasources.jdbc.JDBCOptions.<init>(JDBCOptions.scala:34)
8. at org.apache.spark.sql.execution.datasources.jdbc.JdbcRelationProvider.createRelation(JdbcRelationProvider.scala:32)
9. at org.apache.spark.sql.execution.datasources.DataSource.resolveRelation(DataSource.scala:330)
10. at org.apache.spark.sql.DataFrameReader.load(DataFrameReader.scala:152)
11. at org.apache.spark.sql.DataFrameReader.load(DataFrameReader.scala:125)
12. at com.spark.main.JavaLocalDirectKafkaSparkSQLCarNumber.main(JavaLocalDirectKafkaSparkSQLCarNumber.java:117)
13. at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
14. at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
15. at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
16. at java.lang.reflect.Method.invoke(Method.java:498)
17. at org.apache.spark.deploy.SparkSubmit$.org$apache$spark$deploy$SparkSubmit$$runMain(SparkSubmit.scala:743)
18. at org.apache.spark.deploy.SparkSubmit$.doRunMain$1(SparkSubmit.scala:187)
19. at org.apache.spark.deploy.SparkSubmit$.submit(SparkSubmit.scala:212)
20. at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:126)
21. at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)

1. cd /usr/local/spark/spark-2.1.1-bin-hadoop2.7;
2. bin/spark-submit 
3. --master spark://master:7077 
4. --class "com.spark.main.JavaLocalDirectKafkaSparkSQLCarNumber" 
5. myApp/test-0.0.1-SNAPSHOT-jar-with-dependencies.jar;

需要在spark集群环境上，指定对应的jdbc驱动包：--driver-class-path myApp/ojdbc5.jar

1. cd /usr/local/spark/spark-2.1.1-bin-hadoop2.7;
2. bin/spark-submit 
3. --driver-class-path myApp/ojdbc5.jar 
4. --master spark://master:7077 
5. --class "com.spark.main.JavaLocalDirectKafkaSparkSQLCarNumber" 
6. myApp/test-0.0.1-SNAPSHOT-jar-with-dependencies.jar;

 

再次运行，成功！！！