Spark SQL-对单个查询JAVA中的多列求和

Question

I have over 50 columns for which i want to calculate sum using spark SQL. I don't want to manually write each column name. How can I do it programmatically

Answer 1

Something like

val addNums = df.columns.map(case (c) => df(c))
                .reduce(_ + _)

val sumDF =  df.select(expr(addNums).as("SumOfFifty"))