Group columns by pairs: conditional grouping?

Question

I have a df with three columns: name, test and value. They describe chemical reaction measurements over somebody.

df <- data.frame(name=c("A","A","B","C","C","D"),
                 test=c("urea","creat","urea","urea","creat","creat"),
                 value=c(30,1,40,50,2,0.6))
> df
  name  test value
1    A  urea  30.0
2    A creat   1.0
3    B  urea  40.0
4    C  urea  50.0
5    C creat   2.0
6    D creat   0.6

I need to group or filter only values with both urea or creat test, or equal paired names so df could became:

  name  test value
1    A  urea  30.0
2    A creat   1.0
3    C  urea  50.0
4    C creat   2.0

This is a solution i´ve worked on, but it doesn´t really works.

ndf <- data.frame()
n<-6
while(n>0){
    ifelse(df[n,2]=="creat" && df[n-1,2]=="urea",
        n<-n-2
        ndf <- add_row(ndf, df[(n-1):n,2]),
        n<-n-1)
}

Is there any "tidy" way to group or filter by pairs? maybe row_wise?

Answer 1

I think this is what you want. By each name group, it keeps only groups where the total number of "urea" values is equal to the total number of "creat" values in the test column.

library(dplyr)
df |>
  filter(sum(test == "urea") == sum(test == "creat"), .by = name)