How to fix the loss of attribute sorting / reordering after applying an ML operator?

Mike0985

Dear all,

I have a reordering and sorting of attributes in my process before applying a machine learning model. When I run the process, I will lose the reordering & sorting directly after the machine learning model operator, e.g. Naive Bayes. I also had the same issue with the normalization operator. For now, I put the reorder & sort operator off the process but I still need them.

How can I fix this issue?

Kind regards

Michael

MartinLiebig

Hi Michael,

keep in mind that the order of attributes does not matter in rapidminer. Unlike other tools we identify our columns by name, not by position. The only reason to order them is usually data export. You can then use the Reorder Attributes operator for that.

Best,

Martin

BalazsBaranyRM

Hi Michael,

if you need the data in the original order and attribute order, create a copy with Multiply and then join the later results to that table. This should mostly keep the data in order and add the new columns at the end.

Regards,
Balázs

MartinLiebig

Hi Michael,

keep in mind that the order of attributes does not matter in rapidminer. Unlike other tools we identify our columns by name, not by position. The only reason to order them is usually data export. You can then use the Reorder Attributes operator for that.

Best,

Martin

BalazsBaranyRM

Hi Michael,

if you need the data in the original order and attribute order, create a copy with Multiply and then join the later results to that table. This should mostly keep the data in order and add the new columns at the end.

Regards,
Balázs

Mike0985

@mschmitz @BalazsBarany thank you very much :-).