关于R语言time interval及time_array时间数组的技术咨询

Hey there! Let's work through your time interval problem in R. First, let's get your time array formatted properly in code, then walk through key steps to handle intervals—especially that tricky cross-day time jump (like 23:51:47 followed by 00:18:35).

First, Let's Define Your Time Array #

Here's your time data converted into an R character vector for easy work:

time_array <- c(
  "08:05:02", "08:46:08", "09:13:54", "09:51:21", "10:07:31", "11:34:12",
  "11:45:28", "12:18:14", "12:26:05", "12:58:35", "13:11:09", "14:14:29",
  "15:28:47", "15:56:47", "16:22:06", "16:40:15", "18:15:00", "19:01:39",
  "20:36:22", "21:16:27", "21:26:11", "21:43:06", "23:51:47", "00:18:35",
  "01:02:37", "02:08:14", "02:52:09", "04:02:01", "04:37:46", "05:41:58",
  "05:59:22", "07:22:07", "08:32:47", "09:13:23", "09:39:01", "10:19:35",
  "11:53:18", "12:05:53", "12:18:42", "12:33:04", "13:16:19", "13:37:34",
  "13:54:14", "14:31:39", "14:44:46", "15:26:23", "16:03:25", "17:21:44"
)

Raw time strings aren't useful for calculations. We'll use the hms package to handle hour-minute-second formats cleanly:

# Install and load required packages
install.packages(c("hms", "dplyr"))
library(hms)
library(dplyr)

# Convert to hms (hour-minute-second) objects
time_hms <- as_hms(time_array)

The biggest challenge here is times like 00:18:35 coming right after 23:51:47—R won't automatically know this is the next day. We'll add a date component to fix this:

# Start with a base date (pick any date, it just needs to be consistent)
base_date <- as.Date("2024-01-01")
dates <- rep(base_date, length(time_hms))

# Iterate through times: if current time < previous, increment the date
for (i in 2:length(time_hms)) {
  if (time_hms[i] < time_hms[i-1]) {
    dates[i] <- dates[i-1] + 1
  } else {
    dates[i] <- dates[i-1]
  }
}

# Combine date and time into full POSIXct timestamps
full_time <- as.POSIXct(paste(dates, time_hms))

Now full_time has proper timestamps where cross-day times are correctly marked as the next day.

With full timestamps, calculating intervals between consecutive times is straightforward:

# Calculate intervals between consecutive times (returns difftime objects)
time_intervals <- diff(full_time)

# Convert intervals to minutes (use "hours" or "seconds" for other units)
intervals_minutes <- as.numeric(time_intervals, units = "mins")

# Check the first few intervals
head(intervals_minutes)
# Output: ~41.1, 27.77, 37.45, 16.17, 86.68, 11.27

If you don't need full dates (just want to calculate intervals ignoring dates but accounting for cross-day jumps), you can use this shortcut with lubridate:

install.packages("lubridate")
library(lubridate)

# Convert to period objects
time_period <- hms(time_array)

# Calculate intervals in seconds, fix negative values (cross-day) by adding 1 day (86400 secs)
intervals_seconds <- as.numeric(time_period[-1]) - as.numeric(time_period[-length(time_period)])
intervals_seconds[intervals_seconds < 0] <- intervals_seconds[intervals_seconds < 0] + 86400

# Convert to minutes
intervals_minutes <- intervals_seconds / 60

Now that you have intervals, here are some useful things you can do:

• Filter long intervals: Find gaps longer than 60 minutes

  # Get the timestamps where the following interval is >60 mins
  long_gap_starts <- full_time[-length(full_time)][intervals_minutes > 60]
  

• Summarize interval stats: Get mean, median, min/max intervals

  summary(intervals_minutes)
  

• Visualize interval distribution: Use ggplot2 to see how intervals are spread

  install.packages("ggplot2")
  library(ggplot2)
  
  ggplot(data.frame(interval_mins = intervals_minutes), aes(x = interval_mins)) +
    geom_histogram(binwidth = 5, fill = "steelblue", alpha = 0.7) +
    labs(title = "Distribution of Time Intervals", 
         x = "Interval (Minutes)", 
         y = "Count")
  

内容的提问来源于stack exchange，提问作者statistic