Problem to execute matrix multiplication of csv file which contain header.

[Ohelig Pojke]

library(rmr2)
library(rhdfs)
hdfs.init()
rmr.options(backend="hadoop")
Sum = function(.,YY){
  keyval(1,list(Reduce('+',YY)))
}
XtX = values(from.dfs(mapreduce(
  input = "/data/c.csv",input.format=make.input.format("csv",sep=","),
  map=
    function(.,Xi){ 
      Xi=as.matrix(Xi)
      keyval(1,list(t(Xi)%*% Xi))
    },
  reduce = Sum ,
  combine = TRUE)))[[1]]
My this code work perfect with c.csv, and doesn't work with a.csv. Only difference between these two file is one is with header and one is without header. I removed header of a.csv with excel and saved it as c.csv. How to run a.csv in mapreduce by removing header inside the function.
I have uploaded both files a.csv as well as c.csv.

[Antonio Piccolboni]

It's a more than reasonable request but it's harder to implement then it seems. Each mapper process gets to read a portion of a file. Only the first portion contains the headers. Most processes do  not receive the header information. To make it simple, we decided not to support csv files with headers. If anyone wants to contribute this improvement, pull requests are welcome.

[Ohelig Pojke]

Xi<-as.matrix(setNames(Xi, rep(" ", length(Xi)))) # This is one way to eliminate header i tried this one also as well as
colnames(Xi)<-NULL # This alos i tried.
Xi <- matrix(Xi, ncol = ncol(Xi), dimnames = NULL) #This also remove header and give a matrix without header. But also why my code is not working.