Subset data from the latest date prior to a specified date per group-CodePudding

My data are as follows:

df <- structure(list(site = c("A", "A", "A", "A", "B", "B", "B", "B", 
"A", "A", "A", "A", "B", "B", "B", "B"), year = c(2019L, 2019L, 
2019L, 2019L, 2019L, 2019L, 2019L, 2019L, 2020L, 2020L, 2020L, 
2020L, 2020L, 2020L, 2020L, 2020L), date = c("2019-08-12", "2019-08-14", 
"2019-10-16", "2019-10-27", "2019-08-10", "2019-08-13", "2019-10-01", 
"2019-10-16", "2020-08-08", "2020-08-11", "2020-10-14", "2020-10-16", 
"2020-08-07", "2020-08-19", "2020-10-15", "2020-10-20")), 
class = "data.frame", row.names = c(NA, 
-16L))

For each site and each year, I would like to subset a date range with the following conditions:

The range will begin from the latest date prior to August 15th, and will end at the earliest date after October 15th.

Thank you in advance!

CodePudding user response：

Grouped by 'site', year', filter the 'date' that are between the date value that is less than August 15, and the 'date' value just greater than the October 15th. date[date < as.Date(ISOdate(year, 8, 15))], subsets the dates which are less than the August 15, wrapping with max returns the latest subset date. Similarly for the earliest date is calculated from October 15th.

library(dplyr)
df %>% 
  mutate(date = as.Date(date)) %>%
  group_by(site, year) %>% 
  filter(between(date, max(date[date < as.Date(ISOdate(year, 8, 15))]), 
     min(date[date > as.Date(ISOdate(year, 10, 15))]))) %>%
  ungroup