I have a dataframe, and I am trying to create two new columns, max_start_date and max_end_date, based on a combination of a person id and a date_diff column. My problem is, a person might have multiple date differences. My example is below:
ex <- structure(list(person_id = c(1, 1, 1, 1, 1, 1, 1, 1, 1), serv_from_dt = structure(c(18262,
18262, 18263, 18264, 18275, 18275, 18275, 18278, 18291), class = "Date"),
serv_to_dt = structure(c(18262, 18265, 18263, 18264, 18275,
18278, 18278, 18278, 18291), class = "Date"), days_diff = c(0,
3, 0, 0, 0, 3, 3, 0, 0)), row.names = c(NA,
-9L), class = c("data.table", "data.frame"))
As you can see, the groups of min/max dates are: 2020-01-01/2020-01-04 (days_diff of 3), 2020-01-14/2020-01-17 (days_diff of 3), and 2020-01-30/2020-01-30 (since there are no days overlapping with 2020-01-30).
My desired output looks like so:
output <- structure(list(person_id = c(1, 1, 1, 1, 1, 1, 1, 1, 1), serv_from_dt = structure(c(18262,
18262, 18263, 18264, 18275, 18275, 18275, 18278, 18291), class = "Date"),
serv_to_dt = structure(c(18262, 18265, 18263, 18264, 18275,
18278, 18278, 18278, 18291), class = "Date"), days_diff = c(0,
3, 0, 0, 0, 3, 3, 0, 0), max_start_date = c("2020-01-01",
"2020-01-01", "2020-01-01", "2020-01-01", "2020-01-14", "2020-01-14",
"2020-01-14", "2020-01-14", "2020-01-30"), max_end_date = c("2020-01-04",
"2020-01-04", "2020-01-04", "2020-01-04", "2020-01-17", "2020-01-17",
"2020-01-17", "2020-01-17", "2020-01-30")), row.names = c(NA,
-9L), class = c("data.table", "data.frame"))
As of right now, I have
claims_sample[,days_diff := time_length(serv_to_dt-serv_from_dt, unit = 'days'), prs_nat_key][,`:=`(max_start_date =
serv_from_dt[which.max(days_diff)],
max_end_date = serv_to_dt[which.max(days_diff)]), prs_nat_key]
But this only repeats 2020-01-01 and 2020-01-04 throughout the entire columns. I would really appreciate any help and suggestions on how to solve this. Thanks in advance!
CodePudding user response:
Here is a suggestion: The assumption is that each group of date has 4 dates!
library(dplyr)
ex %>%
group_by(person_id, x = ceiling(row_number()/4)) %>%
mutate(max_start_date = min(serv_from_dt),
max_end_date = max(serv_to_dt)
)
person_id serv_from_dt serv_to_dt days_diff x max_start_date max_end_date
<dbl> <date> <date> <dbl> <dbl> <date> <date>
1 1 2020-01-01 2020-01-01 0 1 2020-01-01 2020-01-04
2 1 2020-01-01 2020-01-04 3 1 2020-01-01 2020-01-04
3 1 2020-01-02 2020-01-02 0 1 2020-01-01 2020-01-04
4 1 2020-01-03 2020-01-03 0 1 2020-01-01 2020-01-04
5 1 2020-01-14 2020-01-14 0 2 2020-01-14 2020-01-17
6 1 2020-01-14 2020-01-17 3 2 2020-01-14 2020-01-17
7 1 2020-01-14 2020-01-17 3 2 2020-01-14 2020-01-17
8 1 2020-01-17 2020-01-17 0 2 2020-01-14 2020-01-17
9 1 2020-01-30 2020-01-30 0 3 2020-01-30 2020-01-30