Delta lake with Java--使用stream同步数据

今天继续学习Delta lake Up and Running 的第8章，处理流数据，要实现的效果就是在一个delta表（名为：YellowTaxiStreamSource）插入一条数据，然后通过流的方式能同步到另外一个delta表 （名为：YellowTaxiStreamTarget）。接着在YellowTaxiStreamSource更新数据YellowTaxiStreamTarget也能更新。至于删除也尝试过了，发现删除是没有办法同步的。

一、先上代码，今天的代码分3份

第1份：用来启动流

import io.delta.tables.DeltaTable;
import org.apache.spark.sql.SparkSession;
import org.apache.spark.sql.streaming.StreamingQueryException;

public class DeltaLakeStream {

    public static void main(String[] args) {
        SparkSession spark = SparkSession.builder()
                .master("local[*]")
                .appName("delta_lake")
                .config("spark.sql.extensions", "io.delta.sql.DeltaSparkSessionExtension")
                .config("spark.sql.catalog.spark_catalog", "org.apache.spark.sql.delta.catalog.DeltaCatalog")
                .getOrCreate();

        String targetPath="D:\\bigdata\\detla-lake-with-java\\YellowTaxiStreamTarget";
        spark.sql("CREATE DATABASE IF NOT EXISTS taxidb");
        //定义源数据表
        spark.sql("CREATE TABLE IF NOT EXISTS taxidb.YellowTaxiSource(" +
                "RideID INT," +
                "PickupTime TIMESTAMP," +
                "CabNumber STRING)" +
                "USING DELTA LOCATION 'file:///D:\\\\bigdata\\\\detla-lake-with-java\\\\YellowTaxiStreamSource'"
        );

        //定义目标数据表
        spark.sql("CREATE TABLE IF NOT EXISTS taxidb.YellowTaxiTarget(" +
                "RideID INT," +
                "PickupTime TIMESTAMP," +
                "CabNumber STRING)" +
                "USING DELTA LOCATION 'file:///D:\\\\bigdata\\\\detla-lake-with-java\\\\YellowTaxiStreamTarget'"
        );

        //通过流的方式读取元数据表，记得要option("ignoreChanges", "true")否则报错
        var stream_df=spark.readStream().option("ignoreChanges", "true").format("delta").load("file:///D:\\\\bigdata\\\\detla-lake-with-java\\\\YellowTaxiStreamSource");

        //打开目标表，用于后面同步数据
        var deltaTable = DeltaTable.forPath(spark, targetPath);
  //   var streamQuery=stream_df.writeStream().format("delta").option("checkpointLocation", targetPath+"\\_checkpoint").start(targetPath);

        //定义同步流，如果目标表的记录与更新记录的RideID相等则更新，没有找到则插入新记录
        var streamQuery=stream_df.writeStream().format("delta").foreachBatch((batchDf,batchId)->{
         deltaTable.as("t").merge(batchDf.as("s"),"t.RideID==s.RideID")
                 .whenMatched().updateAll()
                 .whenNotMatched().insertAll()
                 .execute();


      }).outputMode("Update").start(targetPath);


     try {
            System.out.println("启动stream监听");
            streamQuery.awaitTermination(); //启动流
        } catch (StreamingQueryException e) {
            throw new RuntimeException(e);
        }
    }
}

第2份：用来操作源数据表

import org.apache.spark.sql.SparkSession;

public class DeltaLakeStreamSource {

    public static void main(String[] args) {
        SparkSession spark = SparkSession.builder()
                .master("local[*]")
                .appName("delta_lake")
                .config("spark.sql.extensions", "io.delta.sql.DeltaSparkSessionExtension")
                .config("spark.sql.catalog.spark_catalog", "org.apache.spark.sql.delta.catalog.DeltaCatalog")
                .getOrCreate();


        spark.sql("CREATE DATABASE IF NOT EXISTS taxidb");
        spark.sql("CREATE TABLE IF NOT EXISTS taxidb.YellowTaxiSource(" +
                "RideID INT," +
                "PickupTime TIMESTAMP," +
                "CabNumber STRING)" +
                "USING DELTA LOCATION 'file:///D:\\\\bigdata\\\\detla-lake-with-java\\\\YellowTaxiStreamSource'"
        );
        //验证插入
        spark.sql("INSERT INTO taxidb.YellowTaxiSource (RideID,PickupTime,CabNumber) values (1,'2013-10-13 10:13:15','11-96')").show(false);
        //验证更新  
       //spark.sql("UPDATE taxidb.YellowTaxiSource SET CabNumber='199-99' WHERE RideID=1").show(false);
        //验证删除，不过无效  
        //spark.sql("DELETE FROM taxidb.YellowTaxiSource WHERE RideID=1").show(false);
        spark.sql("SELECT RideID,PickupTime,CabNumber FROM taxidb.YellowTaxiSource").show(false);
        spark.close();
    }
}

第3份：用来验证目标数据表的同步结果

import org.apache.spark.sql.SparkSession;


public class DeltaLakeStreamTarget {

    public static void main(String[] args) {
        SparkSession spark = SparkSession.builder()
                .master("local[*]")
                .appName("delta_lake")
                .config("spark.sql.extensions", "io.delta.sql.DeltaSparkSessionExtension")
                .config("spark.sql.catalog.spark_catalog", "org.apache.spark.sql.delta.catalog.DeltaCatalog")
                .getOrCreate();

        String targetPath="D:\\bigdata\\detla-lake-with-java\\YellowTaxiStreamTarget";
        spark.read().format("delta").load(targetPath).show();

    }
}

二、运行验证

1、先运行DeltaLakeStream，具体运行结果如下图：

2、验证插入数据同步

运行DeltaLakeStreamSource，插入一条RideID=1的数据，具体运行结果如下图：

接着运行 DeltaLakeStreamTarget，看一下数据是否已经通过流的方式同步到目标表，具体运行结果如下图：

3、验证更新数据同步

将DeltaLakeStreamSource的插入数据代码注释掉，同时将更新代码打开，然后运行，将RideID=1的记录的CabNumber值得从11-96修改成199-99，具体运行结果如下图：

接着运行 DeltaLakeStreamTarget，看一下数据是否已经通过流的方式同步到目标表，具体运行结果如下图：

至于删除也尝试过，没有成功，不知道是不是不支持，还望高手指教。