How to summarise and subset multi-level grouped dataframe in dplyr and R

Question

I have the following data in long format:

testdf <- tibble(
          name = c(rep("john", 4), rep("joe", 2)), 
          rep = c(1, 1, 2, 2, 1, 1), 
          field = rep(c("pet", "age"), 3), 
          value = c("dog", "young", "cat", "old", "fish", "young")
)

For each named person (John and Joe), I want to summarise EACH of their pets:
For some reason I can't seem to deal with the repeating events/pets in "John" data.
If I filter just for Joe (only has one pet), the code works.

Any help much appreciated...

testdf %>%
          group_by(name, rep) %>%
        #  filter(name == "joe") %>%  # when I filter only for Joe, the code works
          summarise(
                    about = paste0(
                              "The pet is a: ", .[field == "pet", "value"], " and it is ", .[field == "age", "value"]
                    )
          )

Answer 1

testdf %>%
  pivot_wider(id_cols = name:rep,names_from = field) %>% 
  mutate(about = paste0("The pet is a: ", pet, " and it is ", age))

  name    rep pet   age   about                             
  <chr> <dbl> <chr> <chr> <chr>                             
1 john      1 dog   young The pet is a: dog and it is young 
2 john      2 cat   old   The pet is a: cat and it is old   
3 joe       1 fish  young The pet is a: fish and it is young

This can also be done with data.table, as follows:

library(data.table)

setDT(testdf)[
  ,j = .(about = paste0("The pet is a ", .SD[field=="pet",value], " and it is ", .SD[field=="age",value])),
  by = .(name,rep)
]

   name rep                             about
1: john   1  The pet is a dog and it is young
2: john   2    The pet is a cat and it is old
3:  joe   1 The pet is a fish and it is young

Answer 2

Your data is long format and not tidy, with multiple fields in one. So spread it or pivot wider is what answered by langtang. (better is with data.table but I find it difficult still to use .SD]

I prefer doing these things as simple as possible in dplyr. An alternative -without spreading is as follows which yields same results. [Without data.table where .SD is still difficult for me to grasp! so in 3 lines:

testdf%>%
  group_by(name,rep)%>%    
  summarise(about = paste("The pet is ",value[field=='pet']," and it is ",value[field=='age']))

yields:

      name    rep about                             
  <chr> <dbl> <chr>                             
1 joe       1 The pet is  fish  and it is  young
2 john      1 The pet is  dog  and it is  young 
3 john      2 The pet is  cat  and it is  old