[R-sig-Geo] AttributeList and data.table

pedro at dpi.inpe.br pedro at dpi.inpe.br
Thu Apr 13 17:41:52 CEST 2006


There is a quite new package on CRAN called data.table. It implements 
the class data.table representing a data.frame without rownames, in 
order to improve performance. So, it has the same objective of the sp 
class AttributeList. I confess that I'm very superficial in terms of 
the functionality available in both classes, but I think the projects 
could work together, or even be merged.

Best wishes,

Pedro Andrade

---------- Forwarded message ----------
Date: Wed, 12 Apr 2006 15:19:10 +0100
From: Matthew Dowle <mdowle at concordiafunds.com>
To: "'r-devel at r-project.org'" <r-devel at r-project.org>,
      "'Cran at r-project.org'" <Cran at r-project.org>
Subject: [Rd] New class: data.table


Following previous discussion on this list
(http://tolstoy.newcastle.edu.au/R/devel/05/12/3439.html) I have created a
package as suggested, and uploaded it to CRAN incoming : data.table.tar.gz.

** Your comments and feedback will be very much appreciated. **

> From help(data.table) :

This class really does very little. The only reason for its existence is
that the white book specifies that data.frame must have rownames.

Most of the code is copied from base functions with the code manipulating
row.names removed.

A data.table is identical to a data.frame other than:
  	* it doesn't have rownames
  	* [,drop] by default is FALSE, so selecting a single row will always
return a single row data.table not a vector
  	* The comma is optional inside [], so DT[3] returns the 3rd row as a
1 row data.table
  	* [] is like a call to subset()
  	* [,...], is like a call to with().  (not yet implemented)

  	* up to 10 times less memory
  	* up to 10 times faster to create, and copy
  	* simpler R code
  	* the white book defines rownames, so data.frame can't be changed
... => new class

nr = 1000000
D = rep(1:5,nr/5)
system.time(DF <<- data.frame(colA=D, colB=D))  # 2.08
system.time(DT <<- data.table(colA=D, colB=D))  # 0.15  (over 10 times
faster to create)
identical(as.data.table(DF), DT)
object.size(DF)/object.size(DT)                 # 10 times less memory

tt = subset(DF,colA>3)
ss = DT[colA>3]
identical(as.data.table(tt), ss)


tt = with(subset(DF,colA>3),colA+colB)
ss = with(DT[colA>3],colA+colB)                 # but could be:
DT[colA>3,colA+colB]  (not yet implemented)
identical(tt, ss)

tt = DF[with(DF,tapply(1:nrow(DF),colB,last)),] # select last row grouping
by colB
ss = DT[tapply(1:nrow(DT),colB,last)]           # but could be:
DT[last,group=colB]  (not yet implemented)
identical(as.data.table(tt), ss)

tt = DF[with(DF,colA %in% Lkp),]
ss = DT[colA %in% Lkp]                        # expressions inside the []
can see objects in the calling frame
identical(as.data.table(tt), ss)

In each case above there is either a space, time, or code brevity advantage
with the data.table.

The motivation for the new class grew from the realization that performance
of data.frames can be improved by removing the rownames.  See here for the
previous discussion


R-devel at r-project.org mailing list

More information about the R-sig-Geo mailing list