Apache Spark SQLContext编写错误

Question

I am a novice for using Spark SQL. I followed the online guide here from DataBricks: https://docs.databricks.com/spark/latest/data-sources/sql-databases.html

I can successfully get a connection to the MySQL instance and also read from it. But I keep getting variations of NoTableFound or NoDatabaseFound errors from Spark SQL. Here is what my entire test class looks like:

import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.sql.Dataset;
import org.apache.spark.sql.SQLContext;
import org.apache.spark.sql.SparkSession;

import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.SQLException;
import java.util.Properties;

public class MySqlConnectionTest {
    private static final String MYSQL_USERNAME = "";
    private static final String MYSQL_PASSWORD = "";
    private static final String MYSQL_HOSTNAME = "";
    private static final String MYSQL_PORT = "";
    private static final String MYSQL_DATABASE = "";
    private static final String MYSQL_URL = "jdbc:mysql://" + MYSQL_HOSTNAME + ":" + MYSQL_PORT + "/" + MYSQL_DATABASE + "?user=" + MYSQL_USERNAME + "&password=" + MYSQL_PASSWORD;

public static void main(String[] args) {
    Properties connectionProperties = new Properties();
    connectionProperties.put("user", MYSQL_USERNAME);
    connectionProperties.put("password", MYSQL_PASSWORD);

    /* First verify we are getting a valid connection!
    try {
        testConnection();
    } catch(Exception e) {
        e.printStackTrace();
    } */

    /*
    * NONE of the writeToSummary methods work! The readFromSummary methods work fine...
    * */
//        writeToSummary(connectionProperties);
//        writeToSummaryV2(connectionProperties);
    writeToSummaryV3(connectionProperties);
}

private static void testConnection() throws ClassNotFoundException, SQLException {
    Class.forName("com.mysql.jdbc.Driver");
    Connection connection = DriverManager.getConnection(MYSQL_URL, MYSQL_USERNAME, MYSQL_PASSWORD);
    boolean result = connection.isClosed();
    System.out.println("@@ is connection closed?? ==> " + result);
}

private static SparkSession getSparkSession(){
    return SparkSession.builder().master("local[2]").appName("readUsageSummaryV2").getOrCreate();
}

private static SQLContext getSqlContext() {
    SparkConf sparkConf = new SparkConf()
            .setAppName("saveUsageSummary")
            .setMaster("local[2]");

    JavaSparkContext javaSparkContext = new JavaSparkContext(sparkConf);
    return new SQLContext(javaSparkContext);
}

private static void readFromSummary(Properties connectionProperties) {
    Dataset dataSet = getSqlContext().read().jdbc(MYSQL_URL, "summary", connectionProperties);
    dataSet.printSchema();

    dataSet.select("id","cycle_key", "product", "access_method", "billed", "received_date")
            .limit(5)
            .show();
}

private static void readFromSummaryV2(Properties connectionProperties) {
    Dataset dataSet = getSparkSession().read().jdbc(MYSQL_URL, "summary", connectionProperties);
    dataSet.select("id","cycle_key", "product", "access_method", "billed", "received_date")
            .limit(5)
            .show();
}

private static void writeToSummary(Properties connectionProperties) {
    SQLContext sqlContext = getSqlContext();
    sqlContext.tables("usages")
            .write()
//                .mode(SaveMode.Append)
            .jdbc(MYSQL_URL, "summary", connectionProperties);
}

private static void writeToSummaryV2(Properties connectionProperties) {
    SQLContext sqlContext = getSqlContext();
    sqlContext.table("summary")
            .write()
//        .mode(SaveMode.Append)
            .jdbc(MYSQL_URL, "summary", connectionProperties);
}

private static void writeToSummaryV3(Properties connectionProperties) {
    SQLContext sqlContext = getSqlContext();
    sqlContext.sql("SELECT * FROM summary LIMIT 5")
            .write()
//        .mode(SaveMode.Append)
            .jdbc(MYSQL_URL, "summary", connectionProperties);
}

}

Answer 1

The answer is always a simple one... I re-read the documentation with fresh eyes and understood that for it to work, the Dataset.write() method must be writing something which already exists in Spark SQL context. So I can make it write against a Dataset which is created by reading from the database, like so:

private static void writeToSummaryV4(Properties connectionProperties) {
    Dataset summary = getSparkSession().read().jdbc(MYSQL_URL, "summary", connectionProperties);
    summary.select("comp_code","cycle_key", "product", "access_method", "billed", "received_date")
            .limit(5)
            .show();

    summary.write().mode(SaveMode.Append).jdbc(MYSQL_URL, "summary", connectionProperties);
} 

Answer 2

Another simple way to do this is simply pass a Spark Dataset and write it to any database you want, just pass the correct DB connection strings, like this example below, which writes to a MySQL database.

private static void writeToSummaryV4(Dataset summary) {
    summary.write()
            .format("jdbc")
            .option("url", MYSQL_URL)
            .option("dbtable", MYSQL_DATABASE + "." + MYSQL_SUMMARY_TABLE)
            .option("user", MYSQL_USERNAME)
            .option("password", MYSQL_PASSWORD)
            .mode(SaveMode.Append)
            .save();
} 

For me, I need to read something from a Cassandra database and then load it to the MySQL database. So I could easily get the dataset from Cassandra DB like so:

private static Dataset readFromCassandraSummary() {
    return getSparkSession().read()
            .format("org.apache.spark.sql.cassandra")
            .option("keyspace", "usage")
            .option("table", "summary")
            .load();
}