Extracting and storing data from a very large file in R - r

Question

I have a very large DAT file (16 GB). It contains some information of let's say, 1000 customers. This data is sorted like below that the first column is representing the customer IDs:
9909814 246766 0 31/07/2012 7:00 0.03 0 0 0 0
8211675 262537 0 8/04/2013 3:00 0.52 0 0 0 0
However, the data of customers are not stored in an organized way. So, I want to extract the data of each customer and store it in a separate file. (I have a file that contains the customer IDs. )
For just one customer, I wrote the following code that can search through the file and extract data. However, my problem is to how to do this for all the customers when I'm reading this big file into R.
con<-file('D:/CD_INTERVAL_READING.DAT')
open(con)
n=20
nk=100000
B=9909814 #customer ID for customer no.1
customer1 <- read.table(con, sep=",", nrow=1)
for (i in 1:n) {
conn <- read.table(con,sep=",",skip=(i-1)*nk, nrow=nk)
## extracts just those rows that belong to a specific customer ID
temp1 <-conn[conn$V1==B,]
customer1 <-rbind(customer1,temp1)
}
customer1 <- customer1 [-1,]
library(xlsx)
write.xlsx(customer1, "D:/customer1.xlsx")

Answer 1

The optimal solution would probably be to import the data into a proper database but if you really want to split the file into multiple files based on the first token then you can use awk with this one-liner.
awk '/^/ {ofn=$1 ".txt"} ofn {print > ofn}' filetosplit.txt
It works by
/^/ matching start of line
{ofn=$1 ".txt"} sets the ofn variable to the first word (split by white space) with .txt appended.
Print each line to the file set by ofn.
It takes me just under two minutes on my laptop to split a 1 GB file with the same format as you listed above into multiple text files. I have no idea how well that scales or if it's fast enough for you. If you want an R solution you can always wrap it into a system() call ;o)
Addendum:
Oh ... I'm guessing you are on windows based on the path you mentioned. Then you may need to install Cygwin to get awk.