Data.table rocks! Data manipulation the fast way in R

9 comments
I really should make it a habit of using data.table. The speed and simplicity of this R package are astonishing.

Here is a simple example: I have a data frame showing incremental claims development by line of business and origin year. Now I would like add a column with the cumulative claims position for each line of business and each origin year along the development years.

It's one line with data.table! Here it is:
myData[order(dev), cvalue:=cumsum(value), by=list(origin, lob)]
It is even easy to read! Notice also that I don't have to copy the data. The operator ':=' works by reference and is one of the reasons why data.table is so fast.


And it is getting even better. Suppose you want to get the latest claims development position for each line of business and origin year. Again, it is only one line:
latestData <- myData[, .SD[max(dev)] , by=list(origin, lob)]
Oh boy, I should update my ChainLadder package and utilise the power and elegancy of data.table. Many thanks to Matt Dowle and his collaborators for all their fantastic work.

Here is the R code of the examples above:


Session Info

R Under development (unstable) (2012-10-19 r60974)
Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)

locale:
[1] en_GB.UTF-8/en_GB.UTF-8/en_GB.UTF-8/C/en_GB.UTF-8/en_GB.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] lattice_0.20-10  data.table_1.8.4

loaded via a namespace (and not attached):
[1] grid_2.16.0

9 comments :

  1. Lyudmil Antonov27 November 2012 11:27

    The command:

    myData[order(dev), cvalue:=cumsum(value), by=list(origin, lob)]

    gives

    Error in `[.data.table`(myData, order(dev), `:=`(cvalue, cumsum(value)), :

    Combining := in j with by is not yet implemented. Please let maintainer('data.table') know if you are interested in this.

    ReplyDelete
  2. Do you have an old version? Markus used 1.8.4 - see the output of his sessionInfo(). Type "update.packages()" to upgrade.

    ReplyDelete
  3. This is a nice practical example. Thanks for that.

    I needed two changes to first statement to get it to run

    1. "paste" in place of "paste0"

    2. sep="" at the end to avoide a space in the url

    So it looks like this

    url <- paste("http://www.google.com/fusiontables/api/query?",
    "sql=SELECT+*+FROM+1SL7c4TwyI1YxuQELc0R3PjsYC3TwhP3o7k_NZzc",sep="")

    ReplyDelete
  4. Hi Vijay,

    I guess you use an older version of R than I, as paste0 was added to R with version 2.15.0. See the NEWS section in the R Journal for more details: http://journal.r-project.org/archive/2012-1/RJournal_2012-1.pdf

    Cheers

    Markus

    ReplyDelete
  5. Hi Vijay,

    if you don't want upgrade to a newer R version for some reason try adding this to your .Rprofile:

    ## function paste0
    if (!exists("paste0", where = "package:base")) {
    paste0 <- function(...) paste(..., sep = "")
    }

    Cheers
    harald

    ReplyDelete
  6. Thanks for the post Markus. Beside the speed, would you say that data.table package is a good replacement to plyr and reshape package in its fullest sense?

    ReplyDelete
  7. I am afraid, that I know too little about all those packages to form an opinion.

    ReplyDelete
  8. Thanks Markus. Guess, I will fiddle with it myself and check it out.

    ReplyDelete
  9. Nice question. Yes. v1.8.11 implements fast 'melt' and 'dcast' functions (in C). Have a look at benchmarks here: https://gist.github.com/arunsrinivasan/7839891

    ReplyDelete