Allow rows_update() to update duplicate keys in x

[DavisVaughan]

In this case, I expect that the 2 and 3 in df1$y would both be updated to 4.

I would expect that, in rows_update(), the y data frame should have unique keys, but the x data frame should not have that restriction.

library(dplyr)
​
df1 <- tibble(x = c(1, 1, 2), y = c(2, 3, 5))
df2 <- tibble(x = 1, y = 4)
​
# `x` must have unique keys??
rows_update(df1, df2, by = "x")
#> Error: `x` key values are not unique.

Possibly consider how this affects other rows_*() functions

[olivermagnanimous]

I only just discovered rows_update() and was excited because I need this kind of function often. But I was foiled the first time I tried to use it because it doesn't allow duplicate keys.

library(dplyr)

fruity <- tribble(
  ~fruit,    ~taste,
  "apple",   "sweet",
  "orange",  "sour"
)

truth_about_oranges <- tribble(
  ~ fruit,   ~taste,
  "orange",  "sweet"
)

fruity %>% 
  rows_update(truth_about_oranges, by = "fruit")
#> # A tibble: 2 x 2
#>   fruit  taste
#>   <chr>  <chr>
#> 1 apple  sweet
#> 2 orange sweet


fruity_dup <- tribble(
    ~fruit,    ~taste,
    "apple",   "sweet",
    "orange",  "sour",
    "orange",  "bitter"
  )

fruity_dup %>% 
  rows_update(truth_about_oranges, by = "fruit")
#> Matching, by = "fruit"
#> Error: `x` key values are not unique.

^{Created on 2021-05-11 by the reprex package (v1.0.0)}