理解Spark SQL(三）—— Spark SQL程序举例

上一篇说到，在Spark 2.x当中，实际上SQLContext和HiveContext是过时的，相反是采用SparkSession对象的sql函数来操作SQL语句的。使用这个函数执行SQL语句前需要先调用DataFrame的createOrReplaceTempView注册一个临时表，所以关键是先要将RDD转换成DataFrame。实际上，在Spark中实际声明了

type DataFrame = Dataset[Row]

所以，DataFrame是Dataset[Row]的别名。RDD是提供面向低层次的API，而DataFrame/Dataset提供面向高层次的API（适合于SQL等面向结构化数据的场合）。

下面提供一些Spark SQL程序的例子。

例子一：SparkSQLExam.scala

 package bruce.bigdata.spark.example

 import org.apache.spark.sql.Row
 import org.apache.spark.sql.SparkSession
 import org.apache.spark.sql.types._

 object SparkSQLExam {

     case class offices(office:Int,city:String,region:String,mgr:Int,target:Double,sales:Double)

     def main(args: Array[String]) {

         val spark = SparkSession
           .builder
           .appName("SparkSQLExam")
           .getOrCreate()

         runSparkSQLExam1(spark)
         runSparkSQLExam2(spark)

         spark.stop()

     }

     private def runSparkSQLExam1(spark: SparkSession): Unit = {

         import spark.implicits._

         val rddOffices=spark.sparkContext.textFile("/user/hive/warehouse/orderdb.db/offices/offices.txt").map(_.split("\t")).map(p=>offices(p(0).trim.toInt,p(1),p(2),p(3).trim.toInt,p(4).trim.toDouble,p(5).trim.toDouble))
         val officesDataFrame = spark.createDataFrame(rddOffices)

         officesDataFrame.createOrReplaceTempView("offices")
         spark.sql("select city from offices where region='Eastern'").map(t=>"City: " + t(0)).collect.foreach(println)

     }

     private def runSparkSQLExam2(spark: SparkSession): Unit = {

          import spark.implicits._
          import org.apache.spark.sql._
          import org.apache.spark.sql.types._

          val schema = new StructType(Array(StructField("office", IntegerType, false), StructField("city", StringType, false), StructField("region", StringType, false), StructField("mgr", IntegerType, true), StructField("target", DoubleType, true), StructField("sales", DoubleType, false)))
          val rowRDD = spark.sparkContext.textFile("/user/hive/warehouse/orderdb.db/offices/offices.txt").map(_.split("\t")).map(p => Row(p(0).trim.toInt,p(1),p(2),p(3).trim.toInt,p(4).trim.toDouble,p(5).trim.toDouble))
          val dataFrame = spark.createDataFrame(rowRDD, schema)

          dataFrame.createOrReplaceTempView("offices2")
          spark.sql("select city from offices2 where region='Western'").map(t=>"City: " + t(0)).collect.foreach(println)

     }

 }

使用下面的命令进行编译：

[root@BruceCentOS4 scala]# scalac SparkSQLExam.scala

在编译之前，需要在CLASSPATH中增加路径：

export CLASSPATH=$CLASSPATH:$SPARK_HOME/jars/*:$(/opt/hadoop/bin/hadoop classpath)

然后打包成jar文件：

[root@BruceCentOS4 scala]# jar -cvf spark_exam_scala.jar bruce

然后通过spark-submit提交程序到yarn集群执行，为了方便从客户端查看结果，这里采用yarn cient模式运行。

[root@BruceCentOS4 scala]# $SPARK_HOME/bin/spark-submit --class bruce.bigdata.spark.example.SparkSQLExam --master yarn --deploy-mode client spark_exam_scala.jar

运行结果截图：

例子二：SparkSQLExam.scala（需要启动hive metastore）

 package  bruce.bigdata.spark.example

 import org.apache.spark.sql.{SaveMode, SparkSession}

 object SparkHiveExam {

     def main(args: Array[String]) {

         val spark = SparkSession
           .builder()
           .appName("Spark Hive Exam")
           .config("spark.sql.warehouse.dir", "/user/hive/warehouse")
           .enableHiveSupport()
           .getOrCreate()

         import spark.implicits._

         //使用hql查看hive数据
         spark.sql("show databases").collect.foreach(println)
         spark.sql("use orderdb")
         spark.sql("show tables").collect.foreach(println)
         spark.sql("select city from offices where region='Eastern'").map(t=>"City: " + t(0)).collect.foreach(println)

         //将hql查询出的数据保存到另外一张新建的hive表
         //找出订单金额超过1万美元的产品
         spark.sql("""create table products_high_sales(mfr_id string,product_id string,description string)
                    ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t' LINES TERMINATED BY '\n' STORED AS TEXTFILE""")
         spark.sql("""select mfr_id,product_id,description
                    from products a inner join orders b
                    on a.mfr_id=b.mfr and a.product_id=b.product
                    where b.amount>10000""").write.mode(SaveMode.Overwrite).saveAsTable("products_high_sales")

         //将HDFS文件数据导入到hive表中
         spark.sql("""CREATE TABLE IF NOT EXISTS offices2 (office int,city string,region string,mgr int,target double,sales double )
                    ROW FORMAT DELIMITED FIELDS TERMINATED BY '\t' LINES TERMINATED BY '\n' STORED AS TEXTFILE""")
         spark.sql("LOAD DATA INPATH '/user/hive/warehouse/orderdb.db/offices/offices.txt' INTO TABLE offices2")

         spark.stop()
     }
 }

使用下面的命令进行编译：

[root@BruceCentOS4 scala]# scalac SparkHiveExam.scala

使用下面的命令打包：

[root@BruceCentOS4 scala]# jar -cvf spark_exam_scala.jar bruce

使用下面的命令运行：

[root@BruceCentOS4 scala]# $SPARK_HOME/bin/spark-submit --class bruce.bigdata.spark.example.SparkHiveExam --master yarn --deploy-mode client spark_exam_scala.jar

程序运行结果：

 

另外上述程序运行后，hive中多了2张表：

例子三：spark_sql_exam.py

 from __future__ import print_function

 from pyspark.sql import SparkSession
 from pyspark.sql.types import *

 if __name__ == "__main__":
     spark = SparkSession \
         .builder \
         .appName("Python Spark SQL exam") \
         .config("spark.some.config.option", "some-value") \
         .getOrCreate()

     schema = StructType([StructField("office", IntegerType(), False), StructField("city", StringType(), False),
         StructField("region", StringType(), False), StructField("mgr", IntegerType(), True),
         StructField("Target", DoubleType(), True), StructField("sales", DoubleType(), False)])

     rowRDD = spark.sparkContext.textFile("/user/hive/warehouse/orderdb.db/offices/offices.txt").map(lambda p: p.split("\t")) \
         .map(lambda p: (int(p[0].strip()), p[1], p[2], int(p[3].strip()), float(p[4].strip()), float(p[5].strip())))

     dataFrame = spark.createDataFrame(rowRDD, schema)
     dataFrame.createOrReplaceTempView("offices")
     spark.sql("select city from offices where region='Eastern'").show()

     spark.stop()

执行命令运行程序：

[root@BruceCentOS4 spark]# $SPARK_HOME/bin/spark-submit --master yarn --deploy-mode client spark_sql_exam.py

程序运行结果：

例子四：JavaSparkSQLExam.java

 package bruce.bigdata.spark.example;

 import java.util.ArrayList;
 import java.util.List;

 import org.apache.spark.api.java.JavaRDD;
 import org.apache.spark.api.java.function.Function;
 import org.apache.spark.api.java.function.MapFunction;
 import org.apache.spark.sql.Dataset;
 import org.apache.spark.sql.Row;
 import org.apache.spark.sql.RowFactory;
 import org.apache.spark.sql.SparkSession;
 import org.apache.spark.sql.types.DataTypes;
 import org.apache.spark.sql.types.StructField;
 import org.apache.spark.sql.types.StructType;
 import org.apache.spark.sql.AnalysisException;

 public class JavaSparkSQLExam {
     public static void main(String[] args) throws AnalysisException {
         SparkSession spark = SparkSession
           .builder()
           .appName("Java Spark SQL exam")
           .config("spark.some.config.option", "some-value")
           .getOrCreate();    

         List<StructField> fields = new ArrayList<>();
         fields.add(DataTypes.createStructField("office", DataTypes.IntegerType, false));
         fields.add(DataTypes.createStructField("city", DataTypes.StringType, false));
         fields.add(DataTypes.createStructField("region", DataTypes.StringType, false));
         fields.add(DataTypes.createStructField("mgr", DataTypes.IntegerType, true));
         fields.add(DataTypes.createStructField("target", DataTypes.DoubleType, true));
         fields.add(DataTypes.createStructField("sales", DataTypes.DoubleType, false));

         StructType schema = DataTypes.createStructType(fields);

         JavaRDD<String> officesRDD = spark.sparkContext()
           .textFile("/user/hive/warehouse/orderdb.db/offices/offices.txt", 1)
           .toJavaRDD();

         JavaRDD<Row> rowRDD = officesRDD.map((Function<String, Row>) record -> {
           String[] attributes = record.split("\t");
           return RowFactory.create(Integer.valueOf(attributes[0].trim()), attributes[1], attributes[2], Integer.valueOf(attributes[3].trim()), Double.valueOf(attributes[4].trim()), Double.valueOf(attributes[5].trim()));
         });

         Dataset<Row> dataFrame = spark.createDataFrame(rowRDD, schema);

         dataFrame.createOrReplaceTempView("offices");
         Dataset<Row> results = spark.sql("select city from offices where region='Eastern'");
         results.collectAsList().forEach(r -> System.out.println(r));

         spark.stop();
     }
 }

编译打包后通过如下命令执行：

[root@BruceCentOS4 spark]# $SPARK_HOME/bin/spark-submit --class bruce.bigdata.spark.example.JavaSparkSQLExam --master yarn --deploy-mode client spark_exam_java.jar

运行结果：

上面是一些关于Spark SQL程序的一些例子，分别采用了Scala/Python/Java来编写的。另外除了这三种语言，Spark还支持R语言编写程序，因为我自己也不熟悉，就不举例了。不管用什么语言，其实API都是基本一致的，主要是采用DataFrame和Dataset的高层次API来调用和执行SQL。使用这些API，可以轻松的将结构化数据转化成SQL来操作，同时也能够方便的操作Hive中的数据。