Calculate availability with dates-CodePudding

I have a tibble and want to compute monthly availability (A), defined as

A = uptime / (uptime downtime),

where (monthly) downtime is end - start, by month and uptime is total time (1 month) - downtime. What is the way to compute monthly availability for the year 2019?

This is the sample data

structure(list(start = structure(c(1550048400, 1558008000, 1558703040, 
1561032000, 1560945660, 1563451200), tzone = "UTC", class = c("POSIXct", 
"POSIXt")), end = structure(c(1550143989, 1558008000, 1558956840, 
1561032000, 1560945660, 1563451200), tzone = "GMT", class = c("POSIXct", 
"POSIXt"))), row.names = c(NA, -6L), class = c("tbl_df", "tbl", 
"data.frame"))

CodePudding user response：

First, you have inconsistent "tzone" attributes, one is "UTC" and the other is "GMT". It's minor (and slightly noisy), so I'll preempt the noise (though no change in the results):

attr(dat$end, "tzone") <- "UTC"

A helper function:

fun <- function(mon1, mon2, x = dat) {
  # if either start/end is between mon1/mon2, include it ...
  tmp <- x[with(x, (start >= mon1 & start < mon2) | (end >= mon1 & end < mon2)),] |>
    # ... but if start-to-end straddles a month begin/end, then truncate it
    transform(
      start = pmax(start, mon1),
      end = pmin(end, mon2)
    )
  data.frame(start = mon1, end = mon2) |>
    transform(downtime = c(sum(with(tmp, as.numeric(end - start, units = "hours"))), 0)[1]) |>
    transform(uptime = as.numeric(mon2 - mon1, units = "hours") - downtime) |>
    transform(A = uptime / ( uptime   downtime))
}

And the work in base R:

months <- seq(as.POSIXct("2019-01-01 00:00:00", tz="UTC"), by="1 month", length.out=12)
months
#  [1] "2019-01-01 UTC" "2019-02-01 UTC" "2019-03-01 UTC" "2019-04-01 UTC" "2019-05-01 UTC" "2019-06-01 UTC" "2019-07-01 UTC"
#  [8] "2019-08-01 UTC" "2019-09-01 UTC" "2019-10-01 UTC" "2019-11-01 UTC" "2019-12-01 UTC"
do.call(rbind, Map(fun, months[-12], months[-1]))
#         start        end downtime   uptime         A
# 1  2019-01-01 2019-02-01   0.0000 744.0000 1.0000000
# 2  2019-02-01 2019-03-01  26.5525 645.4475 0.9604874
# 3  2019-03-01 2019-04-01   0.0000 744.0000 1.0000000
# 4  2019-04-01 2019-05-01   0.0000 720.0000 1.0000000
# 5  2019-05-01 2019-06-01  70.5000 673.5000 0.9052419
# 6  2019-06-01 2019-07-01   0.0000 720.0000 1.0000000
# 7  2019-07-01 2019-08-01   0.0000 744.0000 1.0000000
# 8  2019-08-01 2019-09-01   0.0000 744.0000 1.0000000
# 9  2019-09-01 2019-10-01   0.0000 720.0000 1.0000000
# 10 2019-10-01 2019-11-01   0.0000 744.0000 1.0000000
# 11 2019-11-01 2019-12-01   0.0000 720.0000 1.0000000

CodePudding user response：

If you are trying to calculate the value of 'A' for each month, then the process would be:

sum up all the down time in each month
subtract that from the total time in the month to get the uptime
divide the uptime by the total time in the month

This is possible using the lubridate package:

library(lubridate)
library(dplyr)

   data <- data %>% 
      mutate(downtime = end-start,
             month = format(end, "%Y-%m %b"),
             month_time = ceiling_date(end, 
                                       unit = "months") - floor_date(end, 
                                                                     unit = "months")) %>% 
      group_by(month) %>% 
      summarise(downtime = sum(downtime),
                month_time = month_time[1]) %>% 
      mutate(uptime = month_time - downtime,
             A = as.numeric(uptime) / as.numeric(uptime   downtime))