R连续数天计算

Question

I have the next question. If I have the next dataframe

data<- data.frame(
        Time= c("2018-01-01", "2018-01-02", "2018-01-03", "2018-01-04", "2018-01-05"),
        TEN= c(10,20,11,12,16)
    )

and I want to count the consecutive days that TEN < 15 as a new column using R .

I try with

waves_min <- function(df, prop, min_value, min_days) {
    sum(with(rle(df$Temp > min_temp), values & lengths >= min_days))
}

but it returns a total number of that count, not a value for each row.

Any ideas?

Thanks

Answer 1

dplyr + data.table::rleid :

library(dplyr)
library(data.table)

data %>%
  group_by(ID = data.table::rleid(TEN < 15)) %>%
  mutate(Consec_Days = if_else(TEN < 15, row_number(), 0L))

Output:

# A tibble: 7 x 4
# Groups:   ID [5]
  Time         TEN    ID Consec_Days
  <fct>      <dbl> <int>       <int>
1 2018-01-01    10     1           1
2 2018-01-02    20     2           0
3 2018-01-03    11     3           1
4 2018-01-04    12     3           2
5 2018-01-05    16     4           0
6 2018-01-06    17     4           0
7 2018-01-07    14     5           1

data.table :

setDT(data)
data[, Consec_Days := ifelse(TEN < 15, 1:.N, 0L), by = rleid(TEN < 15)]

Output:

         Time TEN Consec_Days
1: 2018-01-01  10           1
2: 2018-01-02  20           0
3: 2018-01-03  11           1
4: 2018-01-04  12           2
5: 2018-01-05  16           0
6: 2018-01-06  17           0
7: 2018-01-07  14           1

Base R + data.table::rleid :

data$Consec_Days <- with(data, ave(TEN, data.table::rleid(TEN < 15), 
                                   FUN = function(x) ifelse(x < 15, seq_along(x), 0L)))

Output:

        Time TEN Consec_Days
1 2018-01-01  10           1
2 2018-01-02  20           0
3 2018-01-03  11           1
4 2018-01-04  12           2
5 2018-01-05  16           0
6 2018-01-06  17           0
7 2018-01-07  14           1

Data:

data <- data.frame(
  Time= c("2018-01-01", "2018-01-02", "2018-01-03", "2018-01-04", "2018-01-05", "2018-01-06", "2018-01-07"),
  TEN= c(10,20,11,12,16,17,14)
)

I've added more rows to OP's sample data to illustrate that these solutions work for all cases.