CODEWITHSUNDEEP
rdplyrtidyr
Ciaran
Ciaran

Reputation: 15

Summing data into groups based on sample time

I have rainfall data over a series of days. My aim is to group non-zero rainfall values into discrete events based on the gap between the Timestamp for observations.

Currently, I have a dataframe with all non-zero rainfall observations. I want to sum these observations into events. I want a new 'event'/group to be triggered if the timegap between row i and row j is greater than 30 mins. Is this possible using R:dplyr?

Upvotes: 0

Views: 52

Answers (1)

Aurèle
Aurèle

Reputation: 12819

Maybe this could do:

library(dplyr)
df <- tibble(
  time = as.POSIXct(c("2016-03-01 08:33:00", "2016-03-01 08:45:00", "2016-03-01 10:00:00", "2016-03-01 10:07:00")),
  value = c(10, 12, 7, 13)
)
df %>% 
  mutate(timegap = difftime(time, lag(time, default = time[1])),
         event = cumsum(as.numeric(timegap) > 30 * 60))

# # A tibble: 4 × 4
#                  time value   timegap event
#                <dttm> <dbl>    <time> <int>
# 1 2016-03-01 08:33:00    10    0 secs     0
# 2 2016-03-01 08:45:00    12  720 secs     0
# 3 2016-03-01 10:00:00     7 4500 secs     1
# 4 2016-03-01 10:07:00    13  420 secs     1

Upvotes: 1

Related Questions