R devel - Sep 2013 - Unable to execute Java MapReduce (Hadoop) code from R using rJava

Hi All,

I have written a Java MapReduce code that runs on Hadoop. My intention is
to create an R package which will call the Java code and execute the job.
Hence, I have written a similar R function. But when I call this function
from R terminal, the Hadoop job is not running. Its just printing few lines
of warning messages and does nothing further. Here is the execution
scenario:

*> source("mueda.R")*
*> mueda(analysis="eda", input="/user/root/sample1.txt",
output="/user/root/eda_test", columns=c("0", "1"),
columnSeparator=",")*
*log4j:WARN No appenders could be found for logger
(org.apache.hadoop.metrics2.lib.MutableMetricsFactory).*
*log4j:WARN Please initialize the log4j system properly.*
*log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for
more info.*
*>*

The warning messages displayed are common while running the Hadoop job
normally (without R), but then it executes properly. What might be the
cause that I am unable to execute the job from R using rJava and there's
not error message as well?

Below is the R code using rJava and calling the Java code written in the
backend:

*library(rJava)
*
*mueda <- function(analysis = "eda",*
*  input = NULL, *
*  output = NULL, *
*  columns = c(NULL), *
*  columnSeparator = NULL,*
*  cat.cut.off = 50,*
*  percentilePoints = c(0,0.01,0.05,0.1,0.25,0.50,0.75,0.90,0.95,0.99,1),*
*  histogram = FALSE) {*
*
*
* if (is.null(input) || is.null(output) || is.null(columns) ||
is.null(columnSeparator)) {*
* stop("Usage: mueda(<analysis>, <input>, <output>,
<columns>,
<columnSeparator>, [<cat.cut.off>], [<percentilePoints>],
[<histogram>]")*
* }*
*
*
* # Gets the absolute path of the external JARS*
* #pkgPath = paste(system.file(package="muEDA"), "/jars",
sep="")*
* pkgPath = paste("../inst", "/jars", sep="")*
*
*
* # Initializes the JVM specifying the directory where the main Java class
resides:*
* .jinit("pkgPath")*
*  *
* # Adds all the required JAR to the class path:*
* .jaddClassPath(paste(pkgPath, "Eda.jar", sep="/"))*
* .jaddClassPath(paste(pkgPath, "commons-cli-1.2.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "hadoop-hdfs-2.0.0-cdh4.3.0.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "slf4j-log4j12-1.6.1.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "commons-configuration-1.6.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "guava-11.0.2.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath,
"hadoop-mapreduce-client-core-2.0.0-cdh4.3.0.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "commons-lang-2.5.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "hadoop-auth-2.0.0-cdh4.3.0.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "log4j-1.2.17.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "commons-logging-1.1.1.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "hadoop-common-2.0.0-cdh4.3.0.jar",
sep="/"))*
* .jaddClassPath(paste(pkgPath, "slf4j-api-1.6.1.jar",
sep="/"))*
*
*
* # Creates the R object for the main Java class:*
* obj <- .jnew("EDA")*
*
*
* if ((analysis == "eda") || (analysis == "univ")) {*
*
*
* # Concatenating the column names to pass as an argument to Java*
* col = columns[1]*
* for(i in 2:length(columns)) {*
* col = paste(col, columns[i], sep = ",")*
* }*
*
*
* switch (analysis,*
*
*
* # Calls the Java main class with the “return type”, “method name”,
“parameters to pass” to perform EDA*
* eda = .jcall(obj, "V", "edaExecute", c("eda",
input, output, col)),*
* # Calls the Java main class with the “return type”, “method name”,
“parameters to pass” to perform*
* # Univariate Analysis*
* univ = .jcall(obj, "V", "edaExecute", c("univ",
input, output, col)))*
* } else if (analysis == "freq") {*
*
*
* # Calls the Java main class with the “return type”, “method name”,
“parameters to pass” to perform*
* # Frequency Analysis*
* .jcall(obj, "V", "edaExecute", c("freq", input,
output, col))*
* } else if ((analysis != "eda") && (analysis !=
"univ") && (analysis !"freq")) {*
* stop("Please provide either \"eda\" or \"univ\" or
\"freq\" for
<analysis> argument")*
* }*
*}*

Regards,
Gaurav

	[[alternative HTML version deleted]]