Aggregate (like count) on non-sql data (like created example set)

ZiggizagZiggizag MemberPosts:4Newbie
edited April 4 inHelp
Hi,

Is there a way to aggregate without SQL?

Let's say I have a huge MySQL data set, and I map some values to arbitrary "categories" through a join with a "created example set". For example, I assign a value "high", to records where "offer_id" was 1,3,5 or 7, and I assign a value "low" to records where "offer_id" was 2, 4 or 6. Presume I do not want to load the mapping "offer_id" --> "category" into a database at the moment, but I like having it in a handy "created example set".

I have noticed, rather unfortunately, that "category" argument (which is set on "example data set" by the "set meta data" operator) is not visible to "aggregate" operator, so I see no easy way to count "high" and "low" records after join.

The question is: how to aggregate by attributes coming from such an "example data set"?

Here you are my process:

Answers

  • yyhuangyyhuang Administrator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, MemberPosts:363RM Data Scientist
    There are at least two methods. You can join the mapping table to the big table. Orreplacethe offer id with dictionary. Then useaggregateoperator.
    I replace the id in golf data, here is the sample process that you can import into your own RM studio.

    Ziggizag BalazsBarany
Sign InorRegisterto comment.