Site icon R-bloggers

Avoid apply() function in large datasets

[This article was first published on R – Predictive Hacks, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

When we are dealing with large datasets and there is a need to calculate some values like the row/column min/max/rank/mean etc we should avoid the apply function because it takes a lot of time. Instead, we can use the matrixStats package and its corresponding functions. Let’s provide some comparisons.

Example of Minimum value per Row

Assume that we want to get the minimum value of each row from a 500 x 500 matrix. Let’s compare the performance of the apply function from the base package versus the rowMins function from the matrixStats package.

library(matrixStats)
library(microbenchmark)
library(ggplot2)

x <- matrix( rnorm(5000 * 5000), ncol = 5000 )

tm <- microbenchmark(apply(x,1,min),
                     rowMins(x).
                     times = 100L
                    )

tm
 

Unit: milliseconds
             expr      min         lq       mean    median        uq       max neval
 apply(x, 1, min) 981.6283 1034.98050 1078.04485 1065.4163 1107.9962 1327.9284   100
       rowMins(x)  42.1838   43.80065   46.55752   45.2255   47.6249   81.3097   100

As we can see from the output above, the apply function was 23 times slower than the rowMins. Below we represent the violin plot

autoplot(tm)
 

To leave a comment for the author, please follow the link and comment on their blog: R – Predictive Hacks.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.