Error when using initial_time_split with drake

Question

Learning how to use drake with tidymodels.

Something about using rsample's initial_time_split(), rather than just initial_split(), is giving me an error, when I run make(plan). I get the following:

#> > target data
#> > target split_data
#> Error in UseMethod("complement"): no applicable method for 'complement' applied to an object of class "rsplit"

Have really been racking my brain on this one. The function works fine, independently (ie. the following works):

I feel like I am missing something pretty basic.

Here is the full drake process in a single file (so that it is easier to post up on stack overflow).

Thanks in advance for the hints, as to what I'm doing wrong.

library(drake)
library(tidyverse)
library(tidymodels)

###################################################################

generate_data <- function() {
  tibble(x = rnorm(1e5), y = rnorm(1e5))
}


split_the_data <- function(data) {
  
  data %>% 
    initial_time_split()
  
}


fit_model <- function(data) {
  
  summary(lm(y ~ x, data = data))
}



###################################################################


plan <- drake_plan(
  
  data = generate_data(),
  
  split_data = split_the_data(data),
  
  model = fit_model(training(split_data))
  
)


###################################################################


make(plan)

score 1 · Accepted Answer · answered Jul 23 '20 at 13:33

This should now be fixed in the current development version (as of 93d60ef41119defc0432cc95d2dd6787e4a00b14). You can install it with

install.packages("remotes")
remotes::install_github("ropensci/drake")

The error happened because drake calls NROW() on every target (for dynamic branching purposes) and apparently NROW() errors on rsplit objects.

library(tidyverse)
library(tidymodels)
#> ── Attaching packages ─────────────────────────────────────── tidymodels 0.1.1 ──
#> ✓ broom     0.7.0      ✓ recipes   0.1.13
#> ✓ dials     0.0.8      ✓ rsample   0.0.7 
#> ✓ infer     0.5.3      ✓ tune      0.1.1 
#> ✓ modeldata 0.0.2      ✓ workflows 0.1.2 
#> ✓ parsnip   0.1.2      ✓ yardstick 0.0.7
#> ── Conflicts ────────────────────────────────────────── tidymodels_conflicts() ──
#> x scales::discard() masks purrr::discard()
#> x dplyr::filter()   masks stats::filter()
#> x recipes::fixed()  masks stringr::fixed()
#> x dplyr::lag()      masks stats::lag()
#> x yardstick::spec() masks readr::spec()
#> x recipes::step()   masks stats::step()

generate_data <- function() {
    tibble(x = rnorm(1e5), y = rnorm(1e5))
}

split_the_data <- function(data) {
    data %>% 
        initial_time_split()
}

NROW(split_the_data(generate_data()))
#> Error in UseMethod("complement"): no applicable method for 'complement' applied to an object of class "rsplit"

^{Created on 2020-07-23 by the reprex package (v0.3.0)}

Issue tracked in https://github.com/ropensci/drake/issues/1300.

Error when using initial_time_split with drake

1 Answers1