"Filter of highest values?"

listslists Member Posts: 39 Guru
edited June 2019 in Help



I read the doc. but did not find the right answer.


Filter rows with the 5 highest values of an attribute?

Would this be possible possible?

Is there an expression?

 I think it could be some kind of cascading filters or easier with a math expression (max etc.), but have no idea how to do so.

I'm also interested in a unique-filter (no duplicated values of an attribute).


Thank you.


Best Answer

  • Options
    listslists Member Posts: 39 Guru
    Solution Accepted

    Update: got it...had to use another subprocess for the db-actions.


  • Options
    Thomas_OttThomas_Ott RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,761 Unicorn

    If you want the Top 5 values you could use a Sort operator and then a Filter Example Range operator. Sorty for decscending direction and then use 1 to 5 for the Filter Example Range. 

  • Options
    listslists Member Posts: 39 Guru

    Thank you Thomas,


    unfortunately this gives me with parameter first example=1 and last example=2

    240 rows. I also noticed that in my case, operator "Generate id" produces duplicated ids...




    PS: I try to write performance data of your modified example to a Sqlite-db.

    It seems I have to use db-mode 'overwrite'...hm a little strange, since I want to produce an archive.

    I guess RM writes sequentially into the db. So I would need some kind of caching (maybe).

    I'm aware of remember & recall. But how to trigger recall only if validation has finished?

    Would this be macro stuff?




Sign In or Register to comment.