Data can only be called 'df' in the {drake} and {targets} packages

[gorkang]

When using workflow packages {drake} and {targets}, the only way to be able to use lme4::lmer() is if the data is called df. If the data frame used has any other name, you get an error.

I reported this issue to the {targets} developer here ropensci/targets#160 and he suggested to check here.

Please see reproducible examples below.

library(drake)
library(lme4)

plan_OK <- drake_plan(
  df = sleepstudy,
  mod = lme4::lmer("Reaction ~ Days + (Days | Subject)", df))

plan_BAD <- drake_plan(
  DF = sleepstudy,
  mod = lme4::lmer("Reaction ~ Days + (Days | Subject)", DF))

make(plan_OK)
make(plan_BAD)

plan_OK works fine, but plan_BAD gives the following error:

Error: target mod failed.
diagnose(mod)$error$message:
use of NULL environment is defunct
diagnose(mod)$error$calls:
lme4::lFormula(formula = "Reaction ~ Days + (Days | Subject)",
data = DF, control = list("nloptwrap", TRUE, 1e-05, TRUE,
FALSE, list("ignore", "stop", "ignore", "stop", "stop",
"message+drop.cols", "warning", "stop"), list(list(
"warning", 0.002, NULL), list("message", 1e-04),
list("warning", 1e-06)), list()))
lme4:::checkFormulaData(formula, data, checkLHS = control$check.formula.LHS ==
"stop")
lme4:::allvarex(environment(formula))
base::vapply(vvec, varex, NA, env)
lme4:::FUN(X[[i]], ...)
base::exists(v, envir = env, inherits = FALSE)

The same happens in {targets}, a newer package similar in scope to {drake}:

library(targets)
library(lme4)

tar_script({
  library(targets)
  tar_option_set(packages = c("targets", "lme4"))

  tar_pipeline(
    tar_target(df2,  sleepstudy),
    tar_target(model, lme4::lmer(Reaction ~ Days + (Days | Subject), data = df2))
  )
})

tar_make()

If data is called df, it works, otherwise (as in the example above), fails with the error :

Error : 'data' not found, and some variables missing from formula environment .
Error: callr subprocess failed: 'data' not found, and some variables missing from formula environment .

[bbolker]

This actually seems to be resolved already in the development version: try remotes::install_github("lme4/lme4") and see how it goes?

The new version will be released in the fairly near future (i.e., probably a couple of weeks)

[gorkang]

It is indeed solved in the development version! I looked around the closed issues but did not find anything obviously related. Sorry for not checking before. Thanks!