Inverting a join

amotley

Is there a way to invert a join in rapidminer? I am using an inner join. So instead of getting the output to be all the examples that match between the two data sets, you get a list of all the examples that do not match.

Is this possible?

MartinLiebig

Hi,

 

isnt that set minus? Otherwise you can maybe simply use an outer join and filter for missings?

 

~Martin

BalazsBaranyRM

Hi,

 

I blogged about generic/advanced joins just two days ago:

http://datascientist.at/2016/06/generic-joins-in-rapidminer/#english

 

This could be interesting for your application. Your join condition would be "e1 != e2". 

 

Set Minus is not exactly a join as it is one-sided. 

 

You could also use an Outer join, keep both join attributes and then filter out those that have non-missing join fields. 

 

Be careful, this will get you a huge number of resulting rows.