generating a 50M line dataset

require(data.table)

STR.VALS <- c('John, Denver','Michael, Davis','John, Carpenter','Steve, Hopkins','Jena, Malone','Katy, Darabont','Lisa, Minelli','Frank, Johnson',NA,'986188.9999.01')
DT.CHUNK <- 5*10^7
DT.FINAL <- 5*10^7
N.CHUNKS <- DT.FINAL/DT.CHUNK
APPENDED <- FALSE

for (ch in seq(1,N.CHUNKS)) {
  DT <- data.table(col1=c(),col2=c(),col3=c(),col4=c())
  DT[,col1 := sample(STR.VALS,DT.CHUNK,replace=TRUE)]
  DT[,col2 := sample(seq(1,20),DT.CHUNK,replace=TRUE)]
  DT[,col3 := sample(LETTERS,DT.CHUNK,replace=TRUE)]
  DT[,col4 := rnorm(DT.CHUNK)]
  fwrite(DT, 'fname.csv', append=APPENDED, col.names=!APPENDED, row.names=TRUE, scipen=9)
  APPENDED <- TRUE
}