gctip_spark_dataframe_isEmpty.scala

// Initial Implementation: Counting rows to check if there is input data or not
val inputStockData: DataFrame = spark.read.json("/path/to/json/files")
val numInputRows = stockData.count
val isInputDataEmpty = if(numInputRows > 0) false else true

if(!isInputDataEmpty) {
  // process input data
} else {
  // no input data. Skip processing
}

// Implementation v2: instead of counting the rows to check the emptyness, use isEmpty api provided by the Dataset API. This is more performant.
val inputStockData: DataFrame = spark.read.json("/path/to/json/files")
if(!inputStockData.isEmpty) {
  // process input data
} else {
  // no input data. Skip processing
}

// Implementation v3: Scala Idiomatic way using pattern matching

val inputStockData: DataFrame = spark.read.json("/path/to/json/files")
inputStockData.isEmpty match {
 case false =>  // process input data
 case true =>   // no input data. Skip processing
}

// BONUS implementation v3: The api is biased on the emptyness. Does isNotEmpty feel more natural to you? Well, unfortunately there is not such in-built API.
// But, this is where Scala is incredibly powerful and magical. It allows you to add extension method. Let's do that.

implicit class MyDatasetExtenion[T](ds: Dataset[T]) {
 def isNotEmpty: Boolean = !ds.isEmpty 
}

val inputStockData: DataFrame = spark.read.json("/path/to/json/files")
inputStockData.isNotEmpty match {
 case true =>  // process input data
 case false =>   // no input data. Skip processing
}