如何在下一行匹配时在第一次出现后删除行

Question

Working in RI am trying to remove all rows following a change. A business is open for 3 years then closes, the closed flag stays in the table for the following years. I want to remove the 2 extra years keeping only the data for the year it closed. Some locations close and reopen in the same year they should not be changed.

I've tried slice on min date when status = "close" but this will not work because of the locations that reopen.

Sample data

date <- c("2014","2015","2016","2017","2018","2019","2016","2017","2018","2019","2015","2016","2017","2018","2018","2019","2019")
ID <- c("1","1","1","1","1", "1","2","2","2","2","3","3","3","3","3","3", "3")
status <- c("open", "open", "open", "close", "close", "close", "open", "open","open","open","open", "open", "open","close", "open", "close", "open")


start <- data.frame(date, ID, status)

Above I want to remove the 2018 and 2019 for ID = 1

date <- c("2014","2015","2016","2017","2016","2017","2018","2019","2015","2016","2017","2018","2018","2019","2019")
ID <- c("1","1","1","1","2","2","2","2","3","3","3","3","3","3", "3")
status <- c("open", "open", "open", "close", "open", "open","open","open","open", "open", "open","close", "open", "close", "open")


ideal_outcome <- data.frame(date, ID, status)

Answer 1

One way using rleid from data.table is to group_by ID and consecutive runs of status , keep only one row for the group where status = "close" and select all rows for "open" .

library(dplyr)
library(data.table)

start %>%
  group_by(ID, group = rleid(status)) %>%
  slice(if (first(status) == "open") seq_len(n()) else  1L) %>%
  mutate(status = replace(as.character(status), 
                    status == "close", "permanently_closed")) %>%  
  ungroup() %>%
  select(-group)

# A tibble: 15 x 3
#   date  ID    status
#   <fct> <fct> <chr> 
# 1 2014  1     open  
# 2 2015  1     open  
# 3 2016  1     open  
# 4 2017  1     permanently_closed 
# 5 2016  2     open  
# 6 2017  2     open  
# 7 2018  2     open  
# 8 2019  2     open  
# 9 2015  3     open  
#10 2016  3     open  
#11 2017  3     open  
#12 2018  3     permanently_closed 
#13 2018  3     open  
#14 2019  3     permanently_closed 
#15 2019  3     open  

However, you don't really need to import data.table just for one function, behaviour of rleid can be replicated with base rle as well

start %>%
  group_by(ID, group = with(rle(as.character(status)), 
                       rep(seq_along(values), lengths))) %>%
   slice(if (first(status) == "open") seq_len(n()) else  1L) %>%
   ungroup() %>%
   select(-group)

---

Another way to create groups as suggested by @Sotos using factor , diff and cumsum

start %>% 
  group_by(grp = as.numeric(as.factor(status)), 
           grp = cumsum(c(TRUE, diff(grp) != 0))) %>%
  slice(if (first(status) == "open") seq_len(n()) else  1L)