在R语言中将模型输出的小时数数值列表转换为POSIXct或Date格式的技术问询

Got it, let's walk through exactly how to turn your decimal hour values into proper datetime objects that play nice with the bupar library.

First, a quick heads-up: POSIXct and Date types represent absolute datetime points, not just relative durations. Since your data tracks hours elapsed since model startup, you’ll need to define a start timestamp (the exact moment your model began running) as a reference point—this is non-negotiable for converting relative hours to absolute timestamps.

Step 1: Set your model's start time #

Replace the example below with your model’s actual launch datetime (use UTC for consistency, or your preferred time zone):

model_start <- as.POSIXct("2024-05-20 08:30:00", tz = "UTC")

Step 2: Build a conversion function #

This function takes your decimal hour values, converts them to seconds (since POSIXct uses seconds since the Unix epoch), then adds that duration to your start time:

decimal_hours_to_posixct <- function(elapsed_hours, start_datetime) {
  # Convert hours to seconds (1 hour = 3600 seconds)
  elapsed_seconds <- elapsed_hours * 3600
  # Add elapsed time to start point to get absolute timestamp
  start_datetime + elapsed_seconds
}

Step 3: Test with your sample values #

Let’s plug in your examples to verify the conversion works as expected:

# Your raw decimal hour data
elapsed_hours <- c(1.75, 25.5)

# Convert to POSIXct timestamps
timestamps <- decimal_hours_to_posixct(elapsed_hours, model_start)

print(timestamps)
# Output:
# [1] "2024-05-20 10:15:00 UTC" "2024-05-21 09:00:00 UTC"

Perfect—1.75 hours becomes 1 hour 45 minutes post-startup, and 25.5 hours translates to 1 day plus 1 hour 30 minutes later.

Step 4: Convert to Date type (if needed) #

If you only need the date component (no time details), wrap the POSIXct objects with as.Date():

dates <- as.Date(timestamps)
print(dates)
# Output:
# [1] "2024-05-20" "2024-05-21"

Integrating with Bupar #

Once you have your POSIXct timestamps, you can easily create an event log for bupar. Here’s a quick example:

library(bupar)

# Sample event data matching your timestamps
event_data <- data.frame(
  case_id = c("case1", "case1", "case2", "case2"),
  activity = c("initiate", "complete", "initiate", "complete"),
  timestamp = timestamps
)

# Create a bupar-compatible event log
event_log <- eventlog(
  event_data,
  case_id = "case_id",
  activity_id = "activity",
  timestamp = "timestamp"
)

This log will work seamlessly with all of bupar’s process mining functions.

内容的提问来源于stack exchange，提问作者hillwalker101