使用逗号分隔符将Spark DataFrame中的一列拆分为多列

Question

我想从Dataframe中的一列创建多个列，使用JavaSpark中的逗号分隔符。

我在DataFrame中的一列中有一个带逗号的值，并希望使用逗号分隔符将其拆分为多列。我有以下代码：

Dataset<Row> dfreq1 = spark.read().format("json").option("inferSchema", "true")
            .load("new.json");

    dfreq1.show(5, 300);


    dfreq1.createOrReplaceTempView("tempdata");

    Dataset<Row> dfreq2 = dfreq1.sqlContext().sql("select split(names, '|') from tempdata");

    dfreq2.show(5, 300);

输入

+----------------------------+
|                         name|
+-----------------------------+
|ABC1,XYZ1,GDH1,KLN1,JUL1,HAI1|
|ABC2,XYZ2,GDH2,KLN2,JUL2,HAI2|
+-----------------------------+

输出

+-----------------------------+
| Cl1| Cl2| Cl3| Cl3|Cl4 | Cl4|
+-----------------------------+
|ABC1|XYZ1|GDH1|KLN1|JUL1|HAI1|
|ABC2|XYZ2|GDH2|KLN2|JUL2|HAI2|
+-----------------------------+

Answer 1

将本列中的 csv 读入数据集

Dataset<Row> df= spark.read
  .option("header",false)
  .option("inferSchema",true)
  .option("delimiter", ",")
  .csv(originalDF.map(x=>x.getString(0)))

Answer 2

 List<String> schemaList =  Arrays.asList("name","gender","sale_amount","event","age","shop_time");
    Column column = functions.col("value");
    Column linesSplit = functions.split(column,"@#");
    for(int i=0;i<schemaList.size();i++){
        lines = lines.withColumn(schemaList.get(i),linesSplit.getItem(i));
    }

Answer 3

你可以试试这个

scala> var dfd =Seq(("ABC1,XYZ1,GDH1,KLN1,JUL1,HAI1"),("ABC2,XYZ2,GDH2,KLN2,JUL2,HAI2")).toDF("name")

scala> dfd.withColumn("temp", split(col("name"), ",")).select((0 until 6).map(i => col("temp").getItem(i).as(s"col$i")): _* ).show
+----+----+----+----+----+----+
|col0|col1|col2|col3|col4|col5|
+----+----+----+----+----+----+
|ABC1|XYZ1|GDH1|KLN1|JUL1|HAI1|
|ABC2|XYZ2|GDH2|KLN2|JUL2|HAI2|
+----+----+----+----+----+----+

希望这对你有帮助