R - 根据前后值清理数据

Question

I have got a table which is later on divided into multiple intervals based on multiple conditions. In some rare cases, I one or multiple rows which do not fall into the defined interval, so I'd like to preform some extra clean-up in the data.

For each group (name, location), if the row value in stop == 0, I need to count how many of those rows are in the interval. If that less then <3, I need to check how many continous rows are market as stop == 1 above and below the interval with zero value. If the count of values with stop == 1 above & below == 1 then I need to change values in the intervals with zero to 1.

I hope the picture will make it more clear:

df <- read.table(text="name location    stop
John    London  1
John    London  1
John    London  1
John    London  1
John    London  1
John    London  1
John    London  1
John    London  0
John    London  0
John    London  1
John    London  1
John    London  1
John    London  1
John    London  1
John    London  1
John    London  0
John    New_York    0
John    New_York    0
John    New_York    0
John    New_York    1
John    New_York    0
",header  = TRUE, stringsAsFactors = FALSE)

Answer 1

You could iterate over the rows, but it seems that all you want to do is replace all instances of 101 with 111 and 1001 with 1111 in stop . You can do this by turning the stop column to string and then make substitutions using gsub() :

stopString = paste0(df$stop, collapse = "")
stopString = gsub("101","111",stopString)
stopString = gsub("1001","1111",stopString)
df$stop = as.numeric(unlist(strsplit(stopString,"")))
> df
   name location stop
1  John   London    1
2  John   London    1
3  John   London    1
4  John   London    1
5  John   London    1
6  John   London    1
7  John   London    1
8  John   London    1
9  John   London    1
10 John   London    1
11 John   London    1
12 John   London    1
13 John   London    1
14 John   London    1
15 John   London    1
16 John   London    0
17 John New_York    0
18 John New_York    0
19 John New_York    0
20 John New_York    1
21 John New_York    0

---

Edit: grouping by name and location:

df <- read.table(text="name location    stop
John    London  1
John    London  0
John    London  1
John    New_York    0
John    New_York    1
John    New_York    0
John    New_York    0
John    New_York    0
John    New_York    1
John    New_York    0
",header  = TRUE, stringsAsFactors = TRUE)

f <- function(x)
{
  stopString = paste0(x, collapse = "")
  stopString = gsub("101","111",stopString)
  stopString = gsub("1001","1111",stopString)
  as.numeric(unlist(strsplit(stopString,"")))
}

> df %>% dplyr::group_by(name, location) %>%
  dplyr::summarise(stop=stop, s=f(stop))
# A tibble: 10 x 4
# Groups:   name, location [2]
   name  location  stop     s
   <fct> <fct>    <int> <dbl>
 1 John  London       1     1
 2 John  London       0     1
 3 John  London       1     1
 4 John  New_York     0     0
 5 John  New_York     1     1
 6 John  New_York     0     0
 7 John  New_York     0     0
 8 John  New_York     0     0
 9 John  New_York     1     1
10 John  New_York     0     0