两者 0 以及 0E-9 都是一样的，所以我不知道你为什么要把它改成 0.000000000 .
这是一个测试-

val dfInput = spark.range(1).withColumn("Qty", col("id").cast(StringType))
    val processDF = dfInput.withColumn("QtyOut",dfInput.col("Qty").cast("decimal(32,9)"))
    processDF.show(false)
    processDF.printSchema()

    processDF.withColumn("NewQtyOut",format_number(processDF.col("QtyOut"),9)).show()
    processDF.withColumn("NewQtyOut",format_number(processDF.col("QtyOut"),9)).printSchema()

    processDF
      .withColumn("isTrue", when(col("QtyOut").equalTo(0), true).otherwise(false))
      .show(false)

输出-

+---+---+------+
|id |Qty|QtyOut|
+---+---+------+
|0  |0  |0E-9  |
+---+---+------+

root
 |-- id: long (nullable = false)
 |-- Qty: string (nullable = false)
 |-- QtyOut: decimal(32,9) (nullable = true)

+---+---+------+-----------+
| id|Qty|QtyOut|  NewQtyOut|
+---+---+------+-----------+
|  0|  0|  0E-9|0.000000000|
+---+---+------+-----------+

root
 |-- id: long (nullable = false)
 |-- Qty: string (nullable = false)
 |-- QtyOut: decimal(32,9) (nullable = true)
 |-- NewQtyOut: string (nullable = true)

+---+---+------+------+
|id |Qty|QtyOut|isTrue|
+---+---+------+------+
|0  |0  |0E-9  |true  |
+---+---+------+------+

请注意 isTrue 列。

编辑-1（基于评论）

转换 decimal 从科学记数法 scale 我们想要-

def bigDecimalFormatter(x: Double, y: Int): Double =
      BigDecimal(x).setScale(y, BigDecimal.RoundingMode.HALF_UP).toDouble

    val decimalFormatter = udf((decimal: Double, scale: Int) => bigDecimalFormatter(decimal, scale))
    processDF.select(decimalFormatter(col("QtyOut"), lit(9)),
      decimalFormatter(lit(1.1000000453E4), lit(5)))
      .show(false)

输出-

+--------------+--------------------+
|UDF(QtyOut, 9)|UDF(11000.000453, 5)|
+--------------+--------------------+
|0.0           |11000.00045         |
+--------------+--------------------+