`juice()` should always return the same number of rows as the template

[DavisVaughan]

In the case of juice() where no terms are selected, a tibble with the same number of rows as the template should be returned rather than a completely empty tibble with 0 rows and 0 columns.

library(recipes)

rec <- recipe(~ ., data = iris) %>% 
  step_log(Sepal.Width)

prepped_rec <- prep(rec, iris)

juice(prepped_rec, all_predictors())
#> # A tibble: 150 x 5
#>    Sepal.Length Sepal.Width Petal.Length Petal.Width Species
#>           <dbl>       <dbl>        <dbl>       <dbl> <fct>  
#>  1          5.1        1.25          1.4         0.2 setosa 
#>  2          4.9        1.10          1.4         0.2 setosa 
#>  3          4.7        1.16          1.3         0.2 setosa 
#>  4          4.6        1.13          1.5         0.2 setosa 
#>  5          5          1.28          1.4         0.2 setosa 
#>  6          5.4        1.36          1.7         0.4 setosa 
#>  7          4.6        1.22          1.4         0.3 setosa 
#>  8          5          1.22          1.5         0.2 setosa 
#>  9          4.4        1.06          1.4         0.2 setosa 
#> 10          4.9        1.13          1.5         0.1 setosa 
#> # … with 140 more rows

juice(prepped_rec, all_outcomes())
#> # A tibble: 0 x 0

^{Created on 2019-11-18 by the reprex package (v0.3.0.9000)}

This would help with tidymodels/hardhat#95 and be more consistent

[DavisVaughan]

This also makes it more consistent with dplyr

suppressPackageStartupMessages(library(recipes))

rec <- recipe(~ Sepal.Width, iris) %>%
  prep(iris)

# should be 150 rows
juice(rec, all_outcomes())
#> # A tibble: 0 x 0

dplyr::select(iris, dplyr::matches("x"))
#> data frame with 0 columns and 150 rows

^{Created on 2019-11-18 by the reprex package (v0.3.0.9000)}