These instructions assume you are using R 2.14.0 or later, and are intended to guide H2O users who wish to interact with H2O through their R console. The H2O R package manages a companion package containing H2O analysis tools in order to ensure that users are using matching versions of H2O and H2O R. Additionally, it provides the means for users to connect R to their running instance of H2O.
STEP 1
The download package can be obtained by clicking on the button Download H2O at http://0xdata.com/h2o.
Unzip the downloaded H2O zip file.
STEP 2
Start an instance of H2O. For help with this see Getting Started from a Downloaded Zip File
Users should be aware that in order for H2O to successfully run through R, an instance of H2O must also simultaneously be running. If the instance of H2O is stopped, the R program will no longer run, and work done will be lost.
STEP 3
Install the H2O package, and the H2O client package simultaneously by entering a command similar to the following at the prompt: (Depending on the version of H2O downloaded users may need to change “h2o_1.0.3.gz” to match the current file name to reflect a different version.)
>install.packages("<unzipped h2o directory>/R/h2o_1.0.3.tar.gz", repos = NULL, type = "source")
This returns output similar to the following:
* installing *source* package ‘h2o’ ...
** R
** preparing package for lazy loading
** help
*** installing help indices
** building package indices
** testing if installed package can be loaded
* DONE (h2o)
STEP 4
Install dependencies for the R package by calling
>h2o.installDepPkgs()
Which returns
Loading required package: bitops
Loading required package: MASS
Loading required package: cluster
Loading required package: mclust
Package 'mclust' version 4.2
Loading required package: flexmix
Loading required package: lattice
STEP 5
>localH2O = h2o.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = FALSE, promptUpgrade = TRUE)
Which returns the following output
Successfully connected to http://localhost:54321
The correct version H2O client package is already running in R, there is no need to call it independently.
STEP 7
Here is an example of using the above object in an H2O call in R
>irisPath = system.file("extdata", "iris.csv", package="h2o")
>iris.hex = h2o.importFile(localH2O, path = irisPath, key = "iris.hex")
>summary(iris.hex)