Netezza Performance Server

Why should you choose R with NPS?

In-database analytics with R. Using R on NPS for Analytics, offers a solution that allows users to get the best of both worlds, the flexibility and richness of the R language and the scalability and enterprise strength integration of IBM NPS for Analytics. Figure 1 shows the basic architecture of this integration.             

 

• One instance of R is running on each of potentially many clients connecting to the data warehouse.
• There are several instances of R running in the database. R is running on the SPU allowing to access and process data in parallel in each partition.
• Additionally, there is an R instance running on the host. Both the R instances on the server, and the instance on the client are extended by packages that provide the seamless integration between both. Users mostly interact with the client instance of R, just as if working locally.

 

• There are several ways for the R client to interact with the database server.

• For more details on local data structures to use in R along with NPS refer NPS Package for R User's Guide
• Also, simple operations on the database can be invoked through R wrappers, allowing R users who are not familiar with SQL to perform these operations.
• This functionality is described in NPS R Library. Instead of loading the data to the client, the R user can invoke own R functions or transformation on the server, where they are executed either on the host, or in parallel on the data slices.
• This is achieved by transferring the user R code to the server, where it is interpreted by the server-side R instances.
• For more details regarding compilation, registration and execution of your custom R code in database refer the R user's guide mentioned earlier.

 

Note: Downloading, installing, and working with Open Source R and all other required packages is subject to the terms and conditions that are mentioned in the appropriate license files of those packages.

R repository: https://cran.r-project.org/

Git repository: https://github.com/IBM/NPS-utils/tree/master/R/NPS_R

R Studio (for R client): https://rstudio.com/products/rstudio/download/