Since Talend Studio generates Java code in the background when one drag-and-drops Spark Big Data Batch Job components from the Palette to the Design workspace, it invokes Scala Spark code, which is the language Spark is natively written in. Java & Scala are sister languages and have rich JVM compatibility.
Often times the Talend components used in Big Data Batch Jobs involve functional programming, like map reduce filter functions, and defining the schema for DB tables to later be created via the use of Spark DataFrames. Hence, Spark's RDD API is used to give Talend fine-grained control over the execution of each row of data in the operation.
It is important to note that earlier versions of Talend used legacy Spark code and its RDD API to integrate component-based logic, but now Talend seamlessly integrates the DataFrame, DataSource & Spark SQL APIs when initially reading source data from various external storage systems and performing transformations on the data.
