R Language

Non-standard evaluation and standard evaluation

Introduction#

Dplyr and many modern libraries in R use non-standard evaluation (NSE) for interactive programming and standard evaluation (SE) for programming1.

For instance, the summarise() function use non-standard evaluation but relies on the summarise_() which uses standard evaluation.

The lazyeval library makes it easy to turn standard evaluation function into NSE functions.

Examples with standard dplyr verbs

NSE functions should be used in interactive programming. However, when developping new functions in a new package, it’s better to use SE version.

Load dplyr and lazyeval :

library(dplyr)
library(lazyeval)

Filtering

NSE version

filter(mtcars, cyl == 8)
filter(mtcars, cyl < 6)
filter(mtcars, cyl < 6 & vs == 1)

SE version (to be use when programming functions in a new package)

filter_(mtcars, .dots = list(~ cyl == 8))
filter_(mtcars, .dots = list(~ cyl < 6))
filter_(mtcars, .dots = list(~ cyl < 6, ~ vs == 1))

Summarise

NSE version

summarise(mtcars,  mean(disp))
summarise(mtcars,  mean_disp = mean(disp))

SE version

summarise_(mtcars, .dots = lazyeval::interp(~ mean(x), x = quote(disp)))
summarise_(mtcars, .dots = setNames(list(lazyeval::interp(~ mean(x), x = quote(disp))), "mean_disp"))
summarise_(mtcars, .dots = list("mean_disp" = lazyeval::interp(~ mean(x), x = quote(disp))))

Mutate

NSE version

mutate(mtcars, displ_l = disp / 61.0237)

SE version

mutate_(
    .data = mtcars, 
    .dots = list(
        "displ_l" = lazyeval::interp(
                        ~ x / 61.0237, x = quote(disp)
            )
         )
)

This modified text is an extract of the original Stack Overflow Documentation created by the contributors and released under CC BY-SA 3.0 This website is not affiliated with Stack Overflow