Not able to write spark dataset to database using

问题:

I need to write my spark dataset to oracle database table. I am using dataset write method with append mode. But getting analysis exception, when the spark job was triggered on cluster using spark2-submit command.

I have read the json file, flattened it and set into a dataset as abcDataset.

Spark Version - 2 Oracle Database JDBC Driver - oracle.jdbc.driver.OracleDriver Programming Language - Java

Dataset<Row> abcDataset= dataframe.select(col('abc').....{and other columns};


Properties dbProperties = new Properties();
            InputStream is = SparkReader.class.getClassLoader().getResourceAsStream("dbProperties.yaml");
            dbProperties.load(is);
            String jdbcUrl = dbProperties.getProperty("jdbcUrl");
            dbProperties.put("driver","oracle.jdbc.driver.OracleDriver");
            String where = "USER123.PERSON";
            abcDataset.write().format("org.apache.spark.sql.execution.datasources.jdbc.DefaultSource").option("driver", "oracle.jdbc.driver.OracleDriver").mode("append").jdbc(jdbcUrl, where, dbProperties);

Expected - to write into database but getting the error below -

org.apache.spark.sql.AnalysisException: Multiple sources found for jdbc (org.apache.spark.sql.execution.datasources.jdbc.JdbcRelationProvider, org.apache.spark.sql.execution.datasources.jdbc.DefaultSource), please specify the fully qualified class name.;
    at org.apache.spark.sql.execution.datasources.DataSource$.lookupDataSource(DataSource.scala:670) 

Do we need to set any additional property in spark submit command, as i am running this on cluster, or any step is missing ?

回答1:

You need to use either abcDataset.write.jdbc or abcDataset.write.format("jdbc") when you are writing via jdbc from spark to rdbms.