Alteryx Designer Desktop Discussions

Find answers, ask questions, and share expertise about Alteryx Designer Desktop and Intelligence Suite.
SOLVED

I need an introduction to Revolution R

patrick_digan
17 - Castor
17 - Castor

Good news: I've downloaded Revolution R (RRO and RRE and the Alteryx RRE predictive).

 

Bad news: I don't understand how to leverage the power of Revolution R.

 

Use case: The summarize tool can't group things and then multiply a variable. I used the aggregate function in R to do this. It was slow, so I also built in the ability to summarize so I could test my R tool against the summarize tool itself for summing. The summarize tool was taking less than 2 seconds on my machine with the attached example. When I tested it out with R (before I found out about any Revolutions), it was taking about 40. I thought (perhaps wrongly) that this fancy Revolution R would significantly speed up my R tool. I got Revolution up and running, only to realize that my example took just as long.


So my questions are: 1) Can I speed up the aggregate function in my example using Revolution R (and if so, how)? 2) Can someone explain some of the benefits of Revolution R and how to practically apply them in Alteryx? I'm a newbie here...

 

Thanks in advance!

1 REPLY 1
KaneG
Alteryx Alumni (Retired)

Hi @patrick_digan,

 

There are a couple of things to look at here.

 

Is your Alteryx using the Revolution install? The normal method that would be taken is to install RRE, then install Alteryx. Alteryx will then pick up that Revolution is present and the option will be available in the Help menu to 'Install Revolution R Predictive' or something similar. 

 

What are you doing with it? In general Revolution R speeds up Predictive processes in 2 ways. Parallelisation of cores using MKL, ScaleR packages & by chunking a process using ScaleR package so that the individual chunks can be handed off. So, basically if the process isn't being chunked or using a package for parallelisation then you won't notice the improvement.

 

RRE is not a package that just automatically speeds up R, rather it greatly improves some processes, and this is mostly evident in predictive as there is more calculations. Generally, for an aggregation, I would say that Alteryx is probably just as fast if not faster using the summarise tool. I will note that there are other advantages for RRE other than just speed. (One example I've seen uses the TeradataR package for In-DB analytics via Revolution & Alteryx).

 

Kane

Labels