Evaluate function within package environment without attaching package

Question

Background

I would like to evaluate a set of R function with a package environment without attaching this package
I would like to avoid using package::

Example

Data

Given sample dummy data set:

# Data --------------------------------------------------------------------

tmpCSV <- tempfile(fileext = ".CSV", pattern = "mtcars_data_")
write.csv(x = mtcars[, 1:5], file = tmpCSV, row.names = FALSE)
# Confirm
# readLines(con = tmpCSV)[1]

Reading

`library`

I could read it with use of read_csv function offered within the readr package.

library(readr)
dta <- read_csv(
    file = tmpCSV,
    col_types = cols(
        mpg = col_double(),
        cyl = col_integer(),
        disp = col_integer(),
        hp = col_integer(),
        drat = col_double()
    )
)

`readr::`

It would be also possible to call readr:: functions directly:

# detach("package:readr", unload=TRUE)
dta <- readr::read_csv(
    file = tmpCSV,
    col_types = readr::cols(
        mpg = readr::col_double(),
        cyl = readr::col_integer(),
        disp = readr::col_integer(),
        hp = readr::col_integer(),
        drat = readr::col_double()
    )
)

Problem

I would like to arrive at the same results using eval/evalq (if possible). Desired syntax would resemble:

eval(expr = read_csv(
    file = tmpCSV,
    col_types = cols(
        mpg = col_double(),
        cyl = col_integer(),
        disp = col_integer(),
        hp = col_integer(),
        drat = col_double()
    )
),
# Naturally, the "" bit does not make sense
envir = "package::readr")

Expected error:

Error in read_csv(file = tmpCSV, col_types = cols(mpg = col_double(), : could not find function "read_csv"

Notes

The task is mostly concerned with accessing package functions without loading package and without calling functions directly via :: and :::. Conceptual equivalent would be using with function and referring to data frame columns without $:

with(mtcars, t.test(disp ~ am))

better example:

with(mtcars, mpg[cyl == 8  &  disp > 350])

Couldn't you come up with an example that doesn't require installing the huge BH package? — Roland
@Roland Yes: with(data = asNamespace("dplyr"), expr = summarise(.data = mtcars, "mean_cyl" = mean(cyl))) — Konrad

Roland Roland · Accepted Answer · 2018-04-16T09:01:16

Just use with? I don't understand why you don't want to use ::.

setwd("E:/temp")

tmpCSV <- tempfile(fileext = ".CSV", pattern = "mtcars_data_")
write.csv(x = mtcars[, 1:5], file = tmpCSV, row.names = FALSE)


dta <- readr::read_csv(
  file = tmpCSV,
  col_types = readr::cols(
    mpg = readr::col_double(),
    cyl = readr::col_integer(),
    disp = readr::col_integer(),
    hp = readr::col_integer(),
    drat = readr::col_double()
  )
)

sessionInfo()
#attached base packages:
#  [1] stats     graphics  grDevices datasets  utils     methods   base     

#loaded via a namespace (and not attached):
#[1] readr_1.1.1      compiler_3.4.4   assertthat_0.2.0 R6_2.2.2         cli_1.0.0       
#[6] hms_0.4.2        tools_3.4.4      pillar_1.2.1     rstudioapi_0.7   tibble_1.4.2    
#[11] crayon_1.3.4     Rcpp_0.12.16     utf8_1.1.3       pkgconfig_2.0.1  rlang_0.2.0     
#[16] fortunes_1.5-4 

dtb <- with(asNamespace("readr"), read_csv(
  file = tmpCSV,
  col_types = cols(
    mpg = col_double(),
    cyl = col_integer(),
    disp = col_integer(),
    hp = col_integer(),
    drat = col_double()
  )))
#same happens here

identical(dta, dtb)
#[1] TRUE