java - 如何使用 Spark 数据集 API (Java) 创建数组列

Question

我有一个名为 outLinks 的数据集，它有两列，一个字符串列和一个数组列。如下所示:

+-------+---------------------+
|    url|collect_set(outlinks)|
+-------+---------------------+
| link91|   [link620, link761]|
|link297| [link999, link942...|
|link246| [link623, link605...|
...

我正在尝试向该表添加更多行，其中每个新行都包含一个字符串和一个空列表。 diff 是一个具有一个字符串列的数据集。

outLinks = outLinks.union(
        diff.map(r ->
                 new Tuple2<>(r.getString(0), DataTypes.createArrayType(DataTypes.StringType)),
                 Encoders.tuple(Encoders.STRING(), Encoders.bean(ArrayType.class))).toDF()); 

我尝试以我能想象到的各种方式定义一个空数组/列表。当我像上面那样做时(我使用 ArrayType 类)，我得到以下异常:

Exception in thread "main" java.util.NoSuchElementException: head of empty list
    at scala.collection.immutable.Nil$.head(List.scala:420)
    at scala.collection.immutable.Nil$.head(List.scala:417)
    at org.apache.spark.sql.catalyst.encoders.ExpressionEncoder$$anonfun$5.apply(ExpressionEncoder.scala:121)
    at org.apache.spark.sql.catalyst.encoders.ExpressionEncoder$$anonfun$5.apply(ExpressionEncoder.scala:120)
    at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:234)
    at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:234)
    at scala.collection.immutable.List.foreach(List.scala:381)
    at scala.collection.TraversableLike$class.map(TraversableLike.scala:234)
    at scala.collection.immutable.List.map(List.scala:285)
    at org.apache.spark.sql.catalyst.encoders.ExpressionEncoder$.tuple(ExpressionEncoder.scala:120)
    at org.apache.spark.sql.catalyst.encoders.ExpressionEncoder$.tuple(ExpressionEncoder.scala:186)
    at org.apache.spark.sql.Encoders$.tuple(Encoders.scala:228)
    at org.apache.spark.sql.Encoders.tuple(Encoders.scala)
    at edu.upenn.cis455.pagerank.PageRankTask.run(PageRankTask.java:96)
    at edu.upenn.cis455.pagerank.PageRankTask.main(PageRankTask.java:30)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:497)
    at org.apache.spark.deploy.SparkSubmit$.org$apache$spark$deploy$SparkSubmit$$runMain(SparkSubmit.scala:755)
    at org.apache.spark.deploy.SparkSubmit$.doRunMain$1(SparkSubmit.scala:180)
    at org.apache.spark.deploy.SparkSubmit$.submit(SparkSubmit.scala:205)
    at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:119)
    at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)

当我使用常规 Java 集合时:

outLinks = outLinks.union(
        diff.map(r ->
                 new Tuple2<>(r.getString(0), Collections.emptyList),
                 Encoders.tuple(Encoders.STRING(), Encoders.javaSerialization(List.class))).toDF());

我收到以下异常:

Exception in thread "main" org.apache.spark.sql.AnalysisException: Union can only be performed on tables with the compatible column types. binary <> array<string> at the second column of the second table;;
'Union
:- Aggregate [url#18], [url#18, collect_set(outlinks#1, 0, 0) AS collect_set(outlinks)#99]
:  +- Deduplicate [url#18, outlinks#1], false
:     +- TypedFilter edu.upenn.cis455.pagerank.PageRankTask$$Lambda$15/603456365@713e49c3, interface org.apache.spark.sql.Row, [StructField(url,StringType,true), StructField(outlinks,StringType,true)], createexternalrow(url#18.toString, outlinks#1.toString, StructField(url,StringType,true), StructField(outlinks,StringType,true))
:        +- Project [url#18, outlinks#1]
:           +- Join Inner, (id#15 = storagepage_id#0)
:              :- Relation[id#15,body#16,lastaccessed#17L,url#18] JDBCRelation(pages) [numPartitions=1]
:              +- Relation[storagepage_id#0,outlinks#1] JDBCRelation(storagepage_outlinks) [numPartitions=1]
+- SerializeFromObject [staticinvoke(class org.apache.spark.unsafe.types.UTF8String, StringType, fromString, input[0, scala.Tuple2, true]._1, true) AS value#1771, encodeusingserializer(input[0, scala.Tuple2, true]._2, false) AS _2#1772]
   +- MapElements edu.upenn.cis455.pagerank.PageRankTask$$Lambda$17/2054286321@4bf89d3d, interface org.apache.spark.sql.Row, [StructField(url,StringType,true)], obj#1770: scala.Tuple2
      +- DeserializeToObject createexternalrow(url#18.toString, StructField(url,StringType,true)), obj#1769: org.apache.spark.sql.Row
         +- Except
            :- Project [url#18]
            :  +- Relation[id#15,body#16,lastaccessed#17L,url#18] JDBCRelation(pages) [numPartitions=1]
            +- Project [url#152]
               +- Aggregate [url#152], [url#152, collect_set(outlinks#1, 0, 0) AS collect_set(outlinks)#99]
                  +- Deduplicate [url#152, outlinks#1], false
                     +- TypedFilter edu.upenn.cis455.pagerank.PageRankTask$$Lambda$15/603456365@713e49c3, interface org.apache.spark.sql.Row, [StructField(url,StringType,true), StructField(outlinks,StringType,true)], createexternalrow(url#152.toString, outlinks#1.toString, StructField(url,StringType,true), StructField(outlinks,StringType,true))
                        +- Project [url#152, outlinks#1]
                           +- Join Inner, (id#149 = storagepage_id#0)
                              :- Relation[id#149,body#150,lastaccessed#151L,url#152] JDBCRelation(pages) [numPartitions=1]
                              +- Relation[storagepage_id#0,outlinks#1] JDBCRelation(storagepage_outlinks) [numPartitions=1]

    at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$class.failAnalysis(CheckAnalysis.scala:39)
    at org.apache.spark.sql.catalyst.analysis.Analyzer.failAnalysis(Analyzer.scala:91)
    at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$$anonfun$checkAnalysis$1$$anonfun$apply$13$$anonfun$apply$14.apply(CheckAnalysis.scala:329)
    at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$$anonfun$checkAnalysis$1$$anonfun$apply$13$$anonfun$apply$14.apply(CheckAnalysis.scala:326)
    at scala.collection.immutable.List.foreach(List.scala:381)
    at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$$anonfun$checkAnalysis$1$$anonfun$apply$13.apply(CheckAnalysis.scala:326)
    at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$$anonfun$checkAnalysis$1$$anonfun$apply$13.apply(CheckAnalysis.scala:315)
    at scala.collection.immutable.List.foreach(List.scala:381)
    at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$$anonfun$checkAnalysis$1.apply(CheckAnalysis.scala:315)
    at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$$anonfun$checkAnalysis$1.apply(CheckAnalysis.scala:78)
    at org.apache.spark.sql.catalyst.trees.TreeNode.foreachUp(TreeNode.scala:127)
    at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$class.checkAnalysis(CheckAnalysis.scala:78)
    at org.apache.spark.sql.catalyst.analysis.Analyzer.checkAnalysis(Analyzer.scala:91)
    at org.apache.spark.sql.execution.QueryExecution.assertAnalyzed(QueryExecution.scala:52)
    at org.apache.spark.sql.Dataset$.ofRows(Dataset.scala:66)
    at org.apache.spark.sql.Dataset.withSetOperator(Dataset.scala:2884)
    at org.apache.spark.sql.Dataset.union(Dataset.scala:1656)
    at edu.upenn.cis455.pagerank.PageRankTask.run(PageRankTask.java:95)
    at edu.upenn.cis455.pagerank.PageRankTask.main(PageRankTask.java:29)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:497)
    at org.apache.spark.deploy.SparkSubmit$.org$apache$spark$deploy$SparkSubmit$$runMain(SparkSubmit.scala:755)
    at org.apache.spark.deploy.SparkSubmit$.doRunMain$1(SparkSubmit.scala:180)
    at org.apache.spark.deploy.SparkSubmit$.submit(SparkSubmit.scala:205)
    at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:119)
    at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)

Answer 1

您可以对字符串数组使用spark.implicits().newStringArrayEncoder()。这是示例。

public class SparkSample {
    public static void main(String[] args) {
        SparkSession spark = SparkSession
                .builder()
                .appName("SparkSample")
                .master("local[*]")
                .getOrCreate();

        List<Tuple2<String,String[]>> inputList = new ArrayList<Tuple2<String,String[]>>();
        inputList.add(new Tuple2<String,String[]>("link91",new String[]{"link620","link761"}));
        inputList.add(new Tuple2<String,String[]>("link297",new String[]{"link999","link942"}));
        Dataset<Row> dataset = spark.createDataset(inputList, Encoders.tuple(Encoders.STRING(), spark.implicits().newStringArrayEncoder())).toDF();
        dataset.show(false);    
    }
}