dplyr::lag() problem with many conditions-CodePudding

I'd like to use dplyr::lag() with ifelse using 3 conditions and rules. If factor S then NA, if the factor is D then monitor, but if D repeat continuously and just only continuously three-time or more, them "Combat" I try to do:

         test<-NULL
         test$groups <-c("S","D","D","D","D","S","D","D","S")
         test <- as.data.frame(test)
         test%>%mutate(decision = ifelse(groups  == "D", "Monitor",ifelse(groups  == "D" & lag(groups, 2) == "D","Monitor", ifelse(groups  == "D"  & lag(groups, 3) == "D", "Combat",NA))))
  groups decision
1      S     <NA>
2      D  Monitor
3      D  Monitor
4      D  Monitor
5      D  Monitor
6      S     <NA>
7      D  Monitor
8      D  Monitor
9      S     <NA>

But is wrong, my correct output is:

  groups decision
1      S     <NA>
2      D  Monitor
3      D  Monitor
4      D  Combat
5      D  Combat
6      S     <NA>
7      D  Monitor
8      D  Monitor
9      S     <NA>

Please, any help with it?

CodePudding user response：

library(data.table)
setDT(test)[,id:=rleid(groups)][groups=="D",decision:=fifelse(1:.N>=3, "Combat","Monitor"), by=id][,.(groups,decision)]

Output:

   groups decision
   <char>   <char>
1:      S     <NA>
2:      D  Monitor
3:      D  Monitor
4:      D   Combat
5:      D   Combat
6:      S     <NA>
7:      D  Monitor
8:      D  Monitor
9:      S     <NA>

Or, you can also do like this (but I still incorporate the use of data.table::rleid

test %>% 
  group_by(id:=data.table::rleid(groups)) %>% 
  mutate(decision = case_when(
    groups=="D" & row_number()>=3 ~ "Combat",
    groups=="D" & row_number()<3 ~"Monitor",
    TRUE~as.character(NA))) %>% 
  ungroup() %>%
  select(!id)

# A tibble: 9 x 2
  groups decision
  <chr>  <chr>   
1 S      NA      
2 D      Monitor 
3 D      Monitor 
4 D      Combat  
5 D      Combat  
6 S      NA      
7 D      Monitor 
8 D      Monitor 
9 S      NA