CODEWITHSUNDEEP
rh2o
FirstName LastName
FirstName LastName

Reputation: 1911

Error in R and H2O initialization

I am following the tutorial at revolutionanalytics and in the beginning where the tutorial says

library(h2o)                # Load H2O library  
localH2O = h2o.init()       # initial H2O locl instance

# Upload iris file from the H2O package into the H2O local instance
iris.hex <-  h2o.uploadFile(localH2O, path = system.file("extdata", "iris.csv", 
package="h2o"), key = "iris.hex")

summary(iris.hex)

When I run the commands in R 3.1.0 x64 for windows (configuration given below) with H2O 2.0.0.9 I get the following output:

> library(h2o)                # Load H2O library  
> localH2O = h2o.init()       # initial H2O locl instance
Successfully connected to http://127.0.0.1:54321 
R is connected to H2O cluster:
Error in names(durationVector) = c("days", "hours", "minutes", "seconds",  : 
'names' attribute [5] must be the same length as the vector [0]

iris.hex <-  h2o.uploadFile(localH2O, path = system.file("extdata", "iris.csv", package="h2o"),       
key = "iris.hex")
Error in h2o.uploadFile.FV(object, path, key, parse, header, sep, col.names,  : 
object 'localH2O' not found

I cannot figure out how to solve this. This issue is not there on any forum. I know that the program is trying to assign a vector of different size to an array of different size. But how do I resolve this and get this to work?

R configuration:

platform       x86_64-w64-mingw32          
arch           x86_64                      
os             mingw32                     
system         x86_64, mingw32             
status                                     
major          3                           
minor          1.0                         
year           2014                        
month          04                          
day            10                          
svn rev        65387                       
language       R                           
version.string R version 3.1.0 (2014-04-10)
nickname       Spring Dance

Upvotes: 1

Views: 2624

Answers (2)

0xF
0xF

Reputation: 586

H2O 3.x versions has an option called strict_version_check=False in python to relax the version conflict between client and server. You can use that.

Upvotes: 1

FirstName LastName
FirstName LastName

Reputation: 1911

Figured it out. My cluster was using an old version of H2O and it is supposed to give an error saying "version mismatch". But for some reason, it was not giving that error. Updating the R H2O package and the H2O cluster jar to the latest stable version resolved the issue.

Upvotes: 2

Related Questions