Read Model operator is deprecated from RapidMiner 7.2 and alternative ?
Hi RapidMiner,
I'm using RapidMiner 7.2 now (upgraded from 7.1), may I know what are the operators that have been removed from version 7.2 (e.g. Read Model, Write Model, etc.) and what are the alternative operators for those? Thank you very much for your help!
Best Regards,
Phi Vu
Best Answer
-
MartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,529 RM Data Scientist
There are actually not just two options, but three. But i think only Store/Retrieve fits here.
1. Store/Retrieve
This stores anything which runs through a connection in the repository. This means that the file is physically written to disc or the underlying SQL DB of the Rm Server.
2. Remember/Recall
With Remember you store the object running thorugh a wire in a "process variable". Thus the object is still in memory and not written to disc. If you don't do anything with it the data will be lost after running the process and the garbage collection things it can delete it.
3. Publish to App/Recall from App
This works pretty similar to Remember/Recall, but is session wide. So you can Recall App objects also from other processes within the same session. That way you have a storage which is not deleted after the process. It is only deleted if you you close studio or delete it manually. It is sometimes a very elegant way to have an in-memory state or memory if you are doing things on server, but it also works locally. of course this is a "creative missuse of operators".
~Martin
- Sr. Director Data Solutions, Altair RapidMiner -
Dortmund, Germany0
Answers
Hi phivucs,
i still got read model in my list, but it is know under Legacy Result Access.
Best,
Martin
Dortmund, Germany
I believe that the reason the "read/write model" operators are now under legacy access is because the current versions of the more general I/O operators "store" and "retrieve" will also work with model objects. These are used for storing and retrieving I/O objects with their full metadata in your repository.
Models might also be able to be used with the remember/recall pair, although I am not really sure what the essential difference is between that set and store/retrieve (I always use store/retrieve myself), so perhaps Martin could expand on the difference in usage between those.
Lindon Ventures
Data Science Consulting from Certified RapidMiner Experts
Thanks for the helpful explanation of the difference between store/retrieve and remember/recall! Very important to know the difference given that nothing is permanently saved when using "remember." And I had not thought of using the app storage operators in the way you describe either.
Lindon Ventures
Data Science Consulting from Certified RapidMiner Experts
You are always welcome.
You can also use it on the server to have a kind of in-memory database to speed up some things like webservices or web apps.
Dortmund, Germany