Yet another operator to simplify data preparation with memisc

20 August 2023

The recently published version 0.99.31.6 of the memisc package also contains an %$$% operator that simplifies routine data preparation steps that hitherto would involve calls to the function within(). It is analogous to the operator %$%, which is provided by the "magrittr" package, but is also defined by this package.

These operators are illustrated by the following code examples.

library(magrittr)
library(memisc) 
set.seed(42)

Loading required package: lattice
Loading required package: MASS

Attaching package: ‘memisc’

The following object is masked from ‘package:magrittr’:

    %$%

The following objects are masked from ‘package:stats’:

    contr.sum, contr.treatment, contrasts

The following object is masked from ‘package:base’:

    as.array

Here we create a simple example data frame:

df <- data.frame(a = 1:7, x = rnorm(7))
df

  a          x
1 1  1.3709584
2 2 -0.5646982
3 3  0.3631284
4 4  0.6328626
5 5  0.4042683
6 6 -0.1061245
7 7  1.5115220

The following code creates two new variables b and x.sq in the data frame using within():

df <- within(df,{
    b <- a + 4
    x.sq <- x^2
})
df

  a          x       x.sq  b
1 1  1.3709584 1.87952706  5
2 2 -0.5646982 0.31888402  6
3 3  0.3631284 0.13186224  7
4 4  0.6328626 0.40051508  8
5 5  0.4042683 0.16343288  9
6 6 -0.1061245 0.01126241 10
7 7  1.5115220 2.28469875 11

This is a bit tedious, because we have to write the name of the data frame (i.e. "df") twice. Using the operator %<>% from the magrittr package one needs to write the name of the data frame only once:

df %<>% within({
    b <- a + 4
    x.sq <- x^2
})
df

  a          x       x.sq  b
1 1  1.3709584 1.87952706  5
2 2 -0.5646982 0.31888402  6
3 3  0.3631284 0.13186224  7
4 4  0.6328626 0.40051508  8
5 5  0.4042683 0.16343288  9
6 6 -0.1061245 0.01126241 10
7 7  1.5115220 2.28469875 11

The magrittr package defines an operator %$% that can be used as a shorthand for with():

with(df, mean(x))

[1] 0.5159882

df %$% mean(x)

[1] 0.5159882

Thus it does not seem to be far-fetched to use an analogous shorthand for within() - which is defined in the most recent version of memisc:

df[c("b","x.sq")] <- NULL

df %$$% {
    b <- a + 4
    x.sq <- x^2
}
df

  a          x  b       x.sq
1 1  1.3709584  5 1.87952706
2 2 -0.5646982  6 0.31888402
3 3  0.3631284  7 0.13186224
4 4  0.6328626  8 0.40051508
5 5  0.4042683  9 0.16343288
6 6 -0.1061245 10 0.01126241
7 7  1.5115220 11 2.28469875

Beside being shorter than a call to within(), it results in a data frame (or data set) in which the variables are ordered by their creation - variables created frist, appear first in the resulting data frame.