Variables in rapidminer processes [SOLVED]

kuehn

Hi

We are starting to use rapidminer - it is an awesome framework. However, we came across a small issue:
We want to save temporary results without harddisk or database ressources in order to save calculation time.
We have large data volumes (minimum 30T data sets) and many subprocesses. We want to calculate average values (subtraction of two inputs) for each subprocess and a single overall average for the entire data set.

Is it possible to temporarily store the results in a loop-specific global variable (e.g. a list)?

Best wishes
Uwe

StaryVena

Hi,
you can store sthi variables in one example set and use "Remember" and "Recall" operators to store it in RAM.

Best
Vaclav

[Deleted User]

Hi Uwe,

could you please post the XML of your process (at least a part of it) so that I can get a better understanding of which values you want to compute.

kuehn wrote:

Hi We are starting to use rapidminer - it is an awesome framework. However, we came across a small issue: We want to save temporary results without harddisk or database ressources in order to save calculation time. We have large data volumes (minimum 30T data sets) and many subprocesses. We want to calculate average values (subtraction of two inputs) for each subprocess and a single overall average for the entire data set. Is it possible to temporarily store the results in a loop-specific global variable (e.g. a list)? Best wishes Uwe

Did I get you right, that you want to store the intermediate results within a separate file and add new results to that file after computing them?

Best,
Edin

kuehn

the solution posted by staryvena sounds useful.
i will check if it is practicable in our case.

it works .. thanx a lot