A speed test comparison of plyr, data.table, and dplyr

ssssssspeed_521872450_d085d1e928

Guest post by Jake Russ

For a recent project I needed to make a simple sum calculation on a rather large data frame (0.8 GB, 4+ million rows, and ~80,000 groups). As an avid user of Hadley Wickham’s packages, my first thought was to use plyr. However, the job took plyr roughly 13 hours to complete.

plyr is extremely efficient and user friendly for most problems, so it was clear to me that I was using it for something it wasn’t meant to do, but I didn’t know of any alternative screwdrivers to use.

I asked for some help on the manipulator Google group , and their feedback led me to data.table and dplyr, a new, and still in progress, package project by Hadley.

What follows is a speed comparison of these three packages incorporating all the feedback from the manipulator folks. They found it informative, so Tal asked me to write it up as a reproducible example.

Continue reading “A speed test comparison of plyr, data.table, and dplyr”

data.table version 1.8.1 – now allowed numeric columns and big-number (via bit64) in keys!

This is a guest post written by Branson Owen, an enthusiastic R and data.table user.

Wow, a long time desired feature of data.table finally came true in version 1.8.1! data.table now allowed numeric columns and big number (via bit64) in keys! This is quite a big thing to me and I believe to many other R users too. Now I can hardly think any weakiness of data.table. Oh, did I mention it also started to support character column in the keys (rather than coerce to factor)?

For people who are not familiar with but interested in data.table package, data.table is an enhanced data.frame for high-speed indexing, ordered joins, assignment, grouping and list columns in a short and flexible syntax. You can take a look at some task examples here:

News from datatable-help mailing list:

* New functions chmatch() and %chin%, faster versions of match() and %in% for character vectors. They are about 4 times faster than match() on the example in ?chmatch.

* New function set(DT,i,j,value) allows fast assignment to elements of DT.


M = matrix(1,nrow=100000,ncol=100)
DF = as.data.frame(M)
DT = as.data.table(M)
system.time(for (i in 1:1000) DF[i,1L] <- i) # 591.000s
system.time(for (i in 1:1000) DT[i,V1:=i]) # 1.158s
system.time(for (i in 1:1000) M[i,1L] <- i) # 0.016s
system.time(for (i in 1:1000) set(DT,i,1L,i)) # 0.027s

* Numeric columns (type ‘double’) are now allowed in keys and ad hoc by. Other types which use ‘double’ (such as POSIXct and bit64) can now be fully supported.

For advanced and creative users, it also officially supported list columns awhile ago (rather than support it by accident). For example, your column could be a list of vectors, where each of the vector has different length. This can allow very flexible and creative ways to manipulate data.

The code example below use “function column”, i.e. a list of functions

> DT = data.table(ID=1:4,A=rnorm(4),B=rnorm(4),fn=list(min,max))
> str(DT)
Classes ‘data.table’ and 'data.frame': 4 obs. of 4 variables:
$ ID: int 1 2 3 4
$ A : num -0.7135 -2.5217 0.0265 1.0102
$ B : num -0.4116 0.4032 0.1098 0.0669
$ fn:List of 4
..$ :function (..., na.rm = FALSE)
..$ :function (..., na.rm = FALSE)
..$ :function (..., na.rm = FALSE)
..$ :function (..., na.rm = FALSE)

> DT[,fn[[1]](A,B),by=ID]
ID V1
[1,] 1 -0.71352508
[2,] 2 0.40322625
[3,] 3 0.02648949
[4,] 4 1.01022266

[ref] https://r-forge.r-project.org/tracker/index.php?func=detail&aid=1302&group_id=240&atid=978