只保留不间断的组

Question

I have a kind of time series which is sorted increasingly. Some observations which have no data at first year but have records regularly after no data season. But some of the groups which started to have data, breaks recording data again.

To be understood, I made up a dummy data frame which represents this situation;

set.seed(1453)

data.frame(id = rep(10:15,4)) %>%
group_by(id) %>%
mutate(year=2012:2015) %>%
arrange(year,.by_group = T) %>%
mutate(some_column = sample(c(NA,1),size = 4,replace = T))

the data looks like;

id  year    some_column

10  2012    1
10  2013    NA
10  2014    1
10  2015    NA
11  2012    NA
11  2013    1
11  2014    1
11  2015    NA
12  2012    NA
12  2013    1
12  2014    NA
12  2015    1
13  2012    1
13  2013    NA
13  2014    1
13  2015    1
14  2012    NA
14  2013    NA
14  2014    NA
14  2015    1
15  2012    NA
15  2013    1
15  2014    1
15  2015    1

I want only these observations;

id  year    some_column

14  2012    NA
14  2013    NA
14  2014    NA
14  2015    1
15  2012    NA
15  2013    1
15  2014    1
15  2015    1

Answer 1

Perhaps this works

library(dplyr)
library(data.table)
df1 %>% 
    group_by(id) %>% 
     filter(n_distinct(rleid(some_column)) <=2)  %>%
     ungroup

-output

# A tibble: 8 x 3
     id  year some_column
  <int> <int>       <int>
1    14  2012          NA
2    14  2013          NA
3    14  2014          NA
4    14  2015           1
5    15  2012          NA
6    15  2013           1
7    15  2014           1
8    15  2015           1

data

df1 <- structure(list(id = c(10L, 10L, 10L, 10L, 11L, 11L, 11L, 11L, 
12L, 12L, 12L, 12L, 13L, 13L, 13L, 13L, 14L, 14L, 14L, 14L, 15L, 
15L, 15L, 15L), year = c(2012L, 2013L, 2014L, 2015L, 2012L, 2013L, 
2014L, 2015L, 2012L, 2013L, 2014L, 2015L, 2012L, 2013L, 2014L, 
2015L, 2012L, 2013L, 2014L, 2015L, 2012L, 2013L, 2014L, 2015L
), some_column = c(1L, NA, 1L, NA, NA, 1L, 1L, NA, NA, 1L, NA, 
1L, 1L, NA, 1L, 1L, NA, NA, NA, 1L, NA, 1L, 1L, 1L)),
 class = "data.frame", row.names = c(NA, 
-24L))