对于从一个数组中获取10个数组，每个数组中有10列，方法grouped可以用于拆分组中的所有列：

val columnCount = 100

val values = 1 to columnCount
val rdd = sparkContext.parallelize(Seq(Row.fromSeq(values)))
val columns = (1 to columnCount).map(idx => StructField(name = "Column" + idx, dataType = IntegerType, nullable = true))
val df = spark.createDataFrame(rdd, StructType(columns))

val chunkSize = 10
val chunks = df.columns.grouped(chunkSize).toSeq
val result: Seq[DataFrame] = chunks.map(columns => df.select(columns.head, columns.tail: _*))
result.head.show(false)

字符串
输出量：

+-------+-------+-------+-------+-------+-------+-------+-------+-------+--------+
|Column1|Column2|Column3|Column4|Column5|Column6|Column7|Column8|Column9|Column10|
+-------+-------+-------+-------+-------+-------+-------+-------+-------+--------+
|1      |2      |3      |4      |5      |6      |7      |8      |9      |10      |
+-------+-------+-------+-------+-------+-------+-------+-------+-------+--------+

型