Spark union two dataframes with different columns. val dfs = Seq(df1, df2, df3) dfs.