Permalink
Find file
Fetching contributors…
Cannot retrieve contributors at this time
22 lines (17 sloc) 1.14 KB

Large datasets in R

I learned that R and especially Rstudio can be very helpful for quick plotting and gaining and understanding of the data, along with generating really nice figures (who doesn’t know the fame of ggplot2 library).

In my analysis, I was routinely handling files of the order of 500-1000 MB, containing millions of lines of data, loading which into the memory and plotting was turning out to be too slow on my 8 GB machine (using the read.table function).

A quick search brought me to the post where I discovered the data.table library. The library provides fast access data frame to load files, almost 10 fold faster. And to my surprise, fread really did process my entire file in a couple of seconds. It also shows a handy notification telling you how much of the data has been parsed, so that you are not kept in hell wondering if something is happening or not.

Also, I learned few really nice tricks of plotting (from the experts) which I will probably share in my blog posts at shorts