left_join but treat NA as a wildcard match

Question

Is there a way to have left_join treat any NA values as a wildcard (ie match all).

For example I have a tibble df_x:

df_x<-tribble(
~`Employee Type`, ~`Employee Name`, ~`Employee Department`,
"Manager"     , "Bob"         , "Accounting",
"Junior"      , "Keith"       , "Accounting",
"Manager"     , "Alice"       , "Finance"   ,
"Junior"      , "Robert"      , "Finance"   ,
"Manager"     , "Claire"      , "I.T."     ,
"Senior"      , "Ashley"      , "I.T."     ,
"Junior"      , "Sam"         , "I.T."     ,
"Junior"      , "Joe"         , "I.T."     
)

and another tibble df_y, specifying the salaries for different roles in different departments, where all managers should get 60000 regardless of which department they work for:

df_y<-tribble(
~`Employee Type`, ~`Employee Department`, ~Salary,
"Manager"     , NA                  , 60000  ,
"Senior"      , "I.T."              , 40000  ,
"Junior"      , "I.T."              , 30000  ,
"Junior"      , "Finance"           , 35000  ,
"Junior"      , "Accounting"        , 35000  
)

What I would like is, after a left_join, I would get:

> left_join(df_x,df_y, *magic argument here*)
Joining, by = c("Employee Type", "Employee Department")
# A tibble: 8 x 4
  `Employee Type` `Employee Name` `Employee Department` Salary
  <chr>           <chr>           <chr>                  <dbl>
1 Manager         Bob             Accounting             60000  
2 Junior          Keith           Accounting             35000
3 Manager         Alice           Finance                60000  
4 Junior          Robert          Finance                35000
5 Manager         Claire          I.T.                   60000
6 Senior          Ashley          I.T.                   40000
7 Junior          Sam             I.T.                   30000
8 Junior          Joe             I.T.                   30000

but what I actually get is:

> left_join(df_x,df_y)
Joining, by = c("Employee Type", "Employee Department")
# A tibble: 8 x 4
  `Employee Type` `Employee Name` `Employee Department` Salary
  <chr>           <chr>           <chr>                  <dbl>
1 Manager         Bob             Accounting                NA
2 Junior          Keith           Accounting             35000
3 Manager         Alice           Finance                   NA
4 Junior          Robert          Finance                35000
5 Manager         Claire          I.T.                      NA
6 Senior          Ashley          I.T.                   40000
7 Junior          Sam             I.T.                   30000
8 Junior          Joe             I.T.                   30000

Is there a way to specify "If the department is NA, then apply the lookup anyway". Or am I looking at this the wrong way and using the wrong function for the job? Is there a more "tidyverse" approach to this?

Answer 1

I would join only on type and then filter after:

df_x %>%
  inner_join(df_y, by = "Employee Type", suffix = c("", "_y")) %>%
  filter(is.na(`Employee Department_y`) | `Employee Department` == `Employee Department_y`) %>%
  select(-`Employee Department_y`)

Answer 2

A possible solution:

library(dplyr)

left_join(df_x,df_y) %>% 
  mutate(Salary = ifelse(is.na(Salary),60000,Salary))