Delta lake with Java--利用spark sql操作数据2

一个平凡的IT农民工2024-05-01 19:35

上一篇文章尝试了建库，建表，插入数据，还差删除和更新，所以在这篇文章补充一下，代码很简单，具体如下：

java 复制代码

import org.apache.spark.sql.SaveMode;
import org.apache.spark.sql.SparkSession;

public class DeltaLakeWithSparkSql2 {

    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 sourcePath="D:\\bookcode\\delta-lake-up-and-running-main\\data\\YellowTaxi\\";
          var df = spark.read().format("parquet").load(sourcePath);
          System.out.println("总数据量："+df.count());
          df.write().format("delta").mode(SaveMode.Overwrite).save("file:///D:\\\\bigdata\\\\detla-lake-with-java\\\\YellowTaxi");


          spark.sql("CREATE DATABASE IF NOT EXISTS taxidb");

          spark.sql("CREATE TABLE IF NOT EXISTS taxidb.YellowTaxi USING DELTA LOCATION 'file:///D:\\\\bigdata\\\\detla-lake-with-java\\\\YellowTaxi'");

          spark.sql("DESCRIBE TABLE taxidb.YellowTaxi").show(false);
          spark.sql("SELECT COUNT(*) from  taxidb.YellowTaxi WHERE VendorID>0").show(false);

          spark.sql("SELECT * from  taxidb.YellowTaxi WHERE tpep_pickup_datetime='2021-01-01 00:30:10'").show(false);
          spark.sql("DELETE FROM taxidb.YellowTaxi WHERE tpep_pickup_datetime='2021-01-01 00:30:10'").show(false);
          spark.sql("SELECT * from  taxidb.YellowTaxi WHERE tpep_pickup_datetime='2021-01-01 00:30:10'").show(false);
          spark.sql("DESCRIBE HISTORY taxidb.YellowTaxi").show(false);

          spark.sql("SELECT INPUT_FILE_NAME(), * from  taxidb.YellowTaxi WHERE tpep_pickup_datetime='2022-01-01 00:35:40'").show(false);
          spark.sql("UPDATE taxidb.YellowTaxi SET passenger_count=99 WHERE tpep_pickup_datetime='2022-01-01 00:35:40'").show(false);
          spark.sql("SELECT INPUT_FILE_NAME(), * from  taxidb.YellowTaxi WHERE tpep_pickup_datetime='2022-01-01 00:35:40'").show(false);
          spark.sql("DESCRIBE HISTORY taxidb.YellowTaxi").show(false);

    }
}

里面涉及的数据集YellowTaxi，在如下地址下载，这个也是Delta Lake Up and Runing的配套仓库

GitHub - benniehaelen/delta-lake-up-and-running: Companion repository for the book 'Delta Lake Up and Running'

具体运行结果如下，没有什么需要注意的，纯粹就是输入一下代码，验证一下结果