减去两个数据帧中的列以获得 Spark Scala 中的差异

Question

Having two dataframes with same columns, I would like to create a resultant dataframe with the difference between the columns, having into account that the dataframes have a lot of columns (and rows).

I guess the approach is first doing an inner join, and then do a "WithColumn" with a subtract inside, but I don't know how to do this in an automated way for a lot of columns .

Example:

first dataframe:

Id	col1	col2	col3	...	colXX
1	1.1	1.2	1.6	...	1.8

second dataframe:

Id	col1	col2	col3	...	colXX
1	1.2	1.2	2.1	...	2.1

Expected dataframe:

Id	diff_col1	diff_col2	diff_col3	...	diff_colXX
1	0.1	0.0	0.5	...	0.3

Thanks beforehand!

Answer 1

First prepare the selection of the differences and then apply it to the resulting dataframe from the join.

val selection = 
    df1.columns.diff(Seq("Id"))
       .map(x => (col(s"df1.$x") - col(s"df2.$x")) as s"diff_$x")

val query = 
    df1.as("df1")
       .join(df2.as("df2"), Seq("Id"), "inner")
       .select((Seq(col("df1.Id")) ++ selection):_*)

Notice the alias on the dataframes in the join matching the names being used in the difference calculations.