Renames levels of factor conditional on unique levels of another factor

Question

Lets say I have a dataframe like so:

df<- 
   plantfam        plantsp                  plantcn
   Asteraceae                               fuzzy leaf
   Aster         


        

Answer 1

Here's the code block first, then I'll explain what I did:

temp <- df %>% 
        filter(is.na(plantsp)) %>% 
        group_by(plantfam, plantcn) %>% 
        summarize(plantsp=NA) %>%
        group_by(plantfam) %>%
        mutate(dummy = cumsum(!is.na(plantcn))) %>%
        mutate(plantsp = paste0(plantfam, " morpho", dummy)) %>%
        select(-dummy)

The first thing I suggest is to remove entries that do not require mutation filter(is.na(plantsp)).

Then aggregate redundant entries with group_by(plantfam, plantcn) %>% summarize(plantsp=NA).

I added a dummy variable that counts plantcn in each group with mutate(dummy = cumsum(!is.na(plantcn))).

I use this dummy variable for creating the string you want mutate(plantsp = paste0(plantfam, " morpho", dummy)).

Finally, get rid of the dummy column with select(-dummy).

This is the resulting data frame:

temp
    plantfam        plantcn            plantsp
       <chr>          <chr>              <chr>
1 Asteraceae     fuzzy leaf Asteraceae morpho1
2 Asteraceae non-fuzzy leaf Asteraceae morpho2
3    Poaceae          3vien    Poaceae morpho1

You can add back the entries that already had plantsp names with:

new.df <- df %>% 
          filter(!is.na(plantsp)) %>%
          full_join( ., temp, by = c("plantfam","plantsp","plantcn"))
new.df

NOTE: You will need to do something a little more complicated if you want to keep redundant entries