Spark RDD/Core 编程 API入门系列之map、filter、textFile、cache、对Job输出结果进行升和降序、union、groupByKey、join、reduce、lookup（一）

1、以本地模式实战map和filter

2、以集群模式实战textFile和cache

3、对Job输出结果进行升和降序

4、union

5、groupByKey

6、join

7、reduce

8、lookup

1、以本地模式实战map和filter

以local的方式，运行spark-shell。

spark@SparkSingleNode:~$ cd /usr/local/spark/spark-1.5.2-bin-hadoop2.6/bin
spark@SparkSingleNode:/usr/local/spark/spark-1.5.2-bin-hadoop2.6/bin$ pwd
/usr/local/spark/spark-1.5.2-bin-hadoop2.6/bin
spark@SparkSingleNode:/usr/local/spark/spark-1.5.2-bin-hadoop2.6/bin$ ./spark-shell

从集合中创建RDD，spark中主要提供了两种函数：parallelize和makeRDD，

scala> val rdd = sc.parallelize(List(1,2,3,4,5))
rdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[0] at parallelize at <console>:21

scala> val mappedRDD = rdd.map(2*_)
mappedRDD: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[1] at map at <console>:23

scala> mappedRDD.collect

得到

res0: Array[Int] = Array(2, 4, 6, 8, 10)

scala>

scala> val filteredRDD = mappedRDD.filter(_ > 4)
16/09/26 20:32:29 INFO storage.BlockManagerInfo: Removed broadcast_0_piece0 on localhost:40688 in memory (size: 1218.0 B, free: 534.5 MB)
16/09/26 20:32:30 INFO spark.ContextCleaner: Cleaned accumulator 1
filteredRDD: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[2] at filter at <console>:25

scala> filteredRDD.collect

注意，一般，生产环境和正宗的写法是。

scala> val filteredRDDAgain = sc.parallelize(List(1,2,3,4,5)).map(2 * _).filter(_ > 4).collect

2、以集群模式实战textFile和cache

启动hadoop集群

spark@SparkSingleNode:/usr/local/hadoop/hadoop-2.6.0$ jps
8457 Jps
spark@SparkSingleNode:/usr/local/hadoop/hadoop-2.6.0$ sbin/start-dfs.sh

启动spark集群

spark@SparkSingleNode:/usr/local/spark/spark-1.5.2-bin-hadoop2.6$ sbin/start-all.sh

spark@SparkSingleNode:/usr/local/spark/spark-1.5.2-bin-hadoop2.6/bin$ ./spark-shell --master spark://SparkSingleNode:7077

读取该文件

scala> val rdd = sc.textFile("/README.md")

使用count统计一下该文件的行数

scala> rdd.count

took 7.018386 s

res0: Long = 98

花了时间7.018386 s

通过观察RDD.scala源代码即可知道cache和persist的区别：
 
def persist(newLevel: StorageLevel): this.type = {
　　if (storageLevel != StorageLevel.NONE && newLevel != storageLevel) {
　　　　throw new UnsupportedOperationException( "Cannot change storage level of an RDD after it was already assigned a level")
　　}
　　sc.persistRDD(this)
　　sc.cleaner.foreach(_.registerRDDForCleanup(this))
　　storageLevel = newLevel
　　this
}
/** Persist this RDD with the default storage level (`MEMORY_ONLY`). */
def persist(): this.type = persist(StorageLevel.MEMORY_ONLY)
 
/** Persist this RDD with the default storage level (`MEMORY_ONLY`). */
def cache(): this.type = persist()
 
可知：
1）RDD的cache()方法其实调用的就是persist方法，缓存策略均为MEMORY_ONLY；
2）可以通过persist方法手工设定StorageLevel来满足工程需要的存储级别；
3）cache或者persist并不是action；
附：cache和persist都可以用unpersist来取消

进行缓存

scala> rdd.cache
res1: rdd.type = MapPartitionsRDD[1] at textFile at <console>:21

执行count，使得缓存生效

scala> rdd.count

took 2.055063 s
res2: Long = 98

花了时间 2.055063 s

再执行，count

took 0.583177 s
res3: Long = 98

花了时间 0.583177 s

总结，我们直接基于cache缓存后的数据，计算所消耗时间大大减少。

正在进行中的spark-shell

接着，对上面的RDD，进行wordcount操作

scala> val wordcount = rdd.flatMap(_.split(' ')).map((_,1)).reduceByKey(_+_)
wordcount: org.apache.spark.rdd.RDD[(String, Int)] = ShuffledRDD[4] at reduceByKey at <console>:23

scala> wordcount.collect

通过saveAsTextFile把数据保存起来

res4: Array[(String, Int)] = Array((package,1), (this,1), (Version"](http://spark.apache.org/docs/latest/building-spark.html#specifying-the-hadoop-version),1), (Because,1), (Python,2), (cluster.,1), (its,1), ([run,1), (general,2), (have,1), (pre-built,1), (locally.,1), (locally,2), (changed,1), (sc.parallelize(1,1), (only,1), (several,1), (This,2), (basic,1), (Configuration,1), (learning,,1), (documentation,3), (YARN,,1), (graph,1), (Hive,2), (first,1), (["Specifying,1), ("yarn-client",1), (page](http://spark.apache.org/documentation.html),1), ([params]`.,1), (application,1), ([project,2), (prefer,1), (SparkPi,2), (<http://spark.apache.org/>,1), (engine,1), (version,1), (file,1), (documentation,,1), (MASTER,1), (example,3), (distribution.,1), (are,1), (params,1), (scala>,1), (DataFrames...
scala> wordcount.saveAsTextFile("/result")

只是，仅仅对每行，做了wordcount而已。

3、对Job输出结果进行升和降序

升序

scala> val wordcount = rdd.flatMap(_.split(' ')).map((_,1)).reduceByKey(_+_).map(x => (x._2,x._1)).sortByKey(true).map(x => (x._2,x._1)).saveAsTextFile("/resultAscSorted")

同理，去下载，不多赘述。

变了

scala> val wordcount = rdd.flatMap(_.split(' ')).map((_,1)).reduceByKey(_+_).map(x => (x._2,x._1)).sortBy(true).map(x => (x._2,x._1)).saveAsTextFile("/resultAscSorted")
<console>:23: error: type mismatch;
found : Boolean(true)
required: ((Int, String)) => ?
val wordcount = rdd.flatMap(_.split(' ')).map((_,1)).reduceByKey(_+_).map(x => (x._2,x._1)).sortBy(true).map(x => (x._2,x._1)).saveAsTextFile("/resultAscSorted")
^

scala>

降序

scala> val wordcount = rdd.flatMap(_.split(' ')).map((_,1)).reduceByKey(_+_).map(x => (x._2,x._1)).sortByKey(false).map(x => (x._2,x._1)).saveAsTextFile("/resultDescSorted")

下载，同理

此刻，成功对Job输出结果进行了排序。

4、union

union的使用

scala> val rdd1 = sc.parallelize(List(('a',1),('b',1)))
rdd1: org.apache.spark.rdd.RDD[(Char, Int)] = ParallelCollectionRDD[26] at parallelize at <console>:21

scala> val rdd2 = sc.parallelize(List(('c',1),('d',1)))
rdd2: org.apache.spark.rdd.RDD[(Char, Int)] = ParallelCollectionRDD[27] at parallelize at <console>:21

scala> rdd1 union rdd2
res6: org.apache.spark.rdd.RDD[(Char, Int)] = UnionRDD[28] at union at <console>:26

scala> val result = rdd1 union rdd2
result: org.apache.spark.rdd.RDD[(Char, Int)] = UnionRDD[29] at union at <console>:25

使用collect操作，查看一下执行结果

scala> result.collect

res7: Array[(Char, Int)] = Array((a,1), (b,1), (c,1), (d,1))

5、groupByKey

scala> val wordcount = rdd.flatMap(_.split(' ')).map((_,1)).groupByKey
wordcount: org.apache.spark.rdd.RDD[(String, Iterable[Int])] = ShuffledRDD[32] at groupByKey at <console>:23

scala> wordcount.collect

res8: Array[(String, Iterable[Int])] = Array((package,CompactBuffer(1)), (this,CompactBuffer(1)), (Version"](http://spark.apache.org/docs/latest/building-spark.html#specifying-the-hadoop-version),CompactBuffer(1)), (Because,CompactBuffer(1)), (Python,CompactBuffer(1, 1)), (cluster.,CompactBuffer(1)), (its,CompactBuffer(1)), ([run,CompactBuffer(1)), (general,CompactBuffer(1, 1)), (YARN,,CompactBuffer(1)), (have,CompactBuffer(1)), (pre-built,CompactBuffer(1)), (locally.,CompactBuffer(1)), (locally,CompactBuffer(1, 1)), (changed,CompactBuffer(1)), (sc.parallelize(1,CompactBuffer(1)), (only,CompactBuffer(1)), (several,CompactBuffer(1)), (learning,,CompactBuffer(1)), (basic,CompactBuffer(1)), (first,CompactBuffer(1)), (This,CompactBuffer(1, 1)), (documentation,CompactBuffer(1, 1, 1)), (Confi...
scala>

6、join

概念知识，参考

http://www.cnblogs.com/goforward/p/4748128.html

scala> val rdd1 = sc.parallelize(List(('a',1),('a',2),('b',3),('b',4)))
rdd1: org.apache.spark.rdd.RDD[(Char, Int)] = ParallelCollectionRDD[33] at parallelize at <console>:21

scala> val rdd2 = sc.parallelize(List(('a',5),('a',6),('b',7),('b',8)))
rdd2: org.apache.spark.rdd.RDD[(Char, Int)] = ParallelCollectionRDD[34] at parallelize at <console>:21

scala> rdd1 join rdd2
res9: org.apache.spark.rdd.RDD[(Char, (Int, Int))] = MapPartitionsRDD[37] at join at <console>:26

scala> val result = rdd1 join rdd2
result: org.apache.spark.rdd.RDD[(Char, (Int, Int))] = MapPartitionsRDD[40] at join at <console>:25

scala> result.collect

res10: Array[(Char, (Int, Int))] = Array((b,(3,7)), (b,(3,8)), (b,(4,7)), (b,(4,8)), (a,(1,5)), (a,(1,6)), (a,(2,5)), (a,(2,6)))

scala>

可见，join操作，完全是一个笛卡尔积的操作。

7、reduce

reduce本身啊，在RDD操作里，属于一个action类型的操作，会导致job作业的提交和执行。

scala> val rdd = sc.parallelize(List(1,2,3,4,5))
rdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[41] at parallelize at <console>:21

scala> rdd.reduce(_+_)

res11: Int = 15

8、lookup

scala> val rdd2 = sc.parallelize(List(('a',5),('a',6),('b',7),('b',8)))
rdd2: org.apache.spark.rdd.RDD[(Char, Int)] = ParallelCollectionRDD[42] at parallelize at <console>:21

scala> rdd2.lookup('a')    //返回一个seq， (5, 6) 是把a对应的所有元素的value提出来组成一个seq

res12: Seq[Int] = WrappedArray(5, 6)