Multiple conditions using replace plyr or dplyr in R

Question

Here is a dataset:

> mydat
species section obs doy ranking
   A      A1  b1 123     2.1
   A      A2  b2 135     2.2
   A      A3  b3 147     2.3
   B      A1  b2 124     2.2
   B      A2  b3 132     2.3
   B      A3  b2 145     2.2
   C      A1  b1 120     2.1
   C      A2  b3 133     2.3
   C      A3  b2 137     2.2 

I am trying to code; for each species where obs==b2, if doy of b2 > doy of b3, then ranking=="2.4". If doy of b2 < doy of b3, then ranking=="2.2" (remains the same), so I get this result:

> mydat2
species section obs doy ranking
   A      A1  b1 123     2.1
   A      A2  b2 135     2.2
   A      A3  b3 147     2.3
   B      A1  b2 124     2.2
   B      A2  b3 132     2.3
   B      A3  b2 145     2.4
   C      A1  b1 120     2.1
   C      A2  b3 133     2.3
   C      A3  b2 137     2.4 

I used the package plyr to avoid loops because I find loops difficult to understand. I know a lot of people use dplyr instead of plyr nowadays, so I'd by glad for an answer using either plyr or dplyr. Here is my clumsy try:

require (plyr)
mydat2 <- ddply(.data=mydat,
            .variables=c("species"),
            function(x){
              return(data.frame(replace(x$ranking, x$doy[x$obs=='b2']>x$doy[x$obs=="b3"],2.4)))})

This works, but there is only species and ranking left in the dataset. How can I code this properly to keep the whole dataset with changes in ranking? Thank you for your help.

Answer 1

ASSUMING each species only ever has one doy value for b3, you could get those values as a single table, then join it to the other table to make your comparisons simple, all using dplyr :

library(dplyr)

# get a single doy value for each species
b3values  <- mydat %>% 
  filter(obs == 'b3') %>% 
  group_by(species) %>% 
  summarize(
    # using min(doy) but if there's only one value, you could use any grouping function like sum, max, etc
    b3doy = min(doy)
  )

# join b3values to your original data
mydat2  <- mydat %>% 
  left_join(b3values, by = 'species') %>% 
  mutate(
    # use case_when() to lay out your conditions and the values you want
    ranking = case_when(
        obs == 'b2' & (doy > b3doy) ~ 2.4
      , obs == 'b2' & (doy < b3doy) ~ 2.2
        # default value is to keep the ranking as-is
      , T ~ ranking
    )
  ) %>%
 # optionally, drop the extra b3doy column
 select(-b3doy)