在R中計算會話持續時間

我有一個資料集，其中包含會話 ID、用戶 ID、UNIX 中的時間戳（我使用 lubridate 轉換）和轉換后的時間戳列。

我想添加另一個名為 session_duration（以秒為單位）的列，它是按會話和用戶分組的 max_time 和 min_time 之間的差異。例如，對于會話 #123 和用戶 345，會話持續時間為 [14-06-2022 17:44:37] - [14-06-2022 17:44:32]，即 5 秒。

這就是我當前的代碼的樣子。時間戳已成功轉換，但我遇到了會話持續時間列的問題。

library(tidyverse)
library(lubridate)
df <- df %>%
  mutate(timestamp = as_datetime(ts_unix/1000)) %>%
  group_by (session, user, timestamp) %>%
  mutate(session_duration = difftime (max(timestamp), min(timestamp), units = "secs"))

有人可以幫我弄清楚 session_duration 列嗎？謝謝你。

uj5u.com熱心網友回復：

library(data.table)
setDT(df)[, duration := max(TimeStamp) - min(TimeStamp), by = .(Session)][]

#    Session User        ts_UNIX           TimeStamp  duration
# 1:     123  345 UNIX Timestamp 2022-06-14 17:44:32    5 secs
# 2:     123  345 UNIX Timestamp 2022-06-14 17:44:33    5 secs
# 3:     123  345 UNIX Timestamp 2022-06-14 17:44:37    5 secs
# 4:     124  346 UNIX Timestamp 2022-06-14 15:50:10 2390 secs
# 5:     124  346 UNIX Timestamp 2022-06-14 15:51:01 2390 secs
# 6:     124  346 UNIX Timestamp 2022-06-14 16:30:00 2390 secs
# 7:     125  345 UNIX Timestamp 2022-06-14 23:55:30  329 secs
# 8:     125  345 UNIX Timestamp 2022-06-14 23:58:50  329 secs
# 9:     125  345 UNIX Timestamp 2022-06-14 23:59:45  329 secs
#10:     125  345 UNIX Timestamp 2022-06-15 00:00:32  329 secs
#11:     125  345 UNIX Timestamp 2022-06-15 00:00:59  329 secs

樣本資料

df <- fread("Session    User    ts_UNIX     TimeStamp
123     345     UNIX Timestamp  14-06-2022T17:44:32
123     345     UNIX Timestamp  14-06-2022T17:44:33
123     345     UNIX Timestamp  14-06-2022T17:44:37
124     346     UNIX Timestamp  14-06-2022T15:50:10
124     346     UNIX Timestamp  14-06-2022T15:51:01
124     346     UNIX Timestamp  14-06-2022T16:30:00
125     345     UNIX Timestamp  14-06-2022T23:55:30
125     345     UNIX Timestamp  14-06-2022T23:58:50
125     345     UNIX Timestamp  14-06-2022T23:59:45
125     345     UNIX Timestamp  15-06-2022T00:00:32
125     345     UNIX Timestamp  15-06-2022T00:00:59")

df[, TimeStamp := as.POSIXct(TimeStamp, format= "%d-%m-%YT%H:%M:%S")]

uj5u.com熱心網友回復：

library(tidyverse)
library(lubridate)

df %>% 
  group_by(Session, User) %>% 
  mutate(session_duration = max(TimeStamp) - min(TimeStamp))

# A tibble: 11 × 5
# Groups:   Session, User [3]
   Session  User ts_UNIX        TimeStamp           session_duration
     <dbl> <dbl> <chr>          <dttm>              <drtn>          
 1     123   345 UNIX Timestamp 2022-06-14 17:44:32    5 secs       
 2     123   345 UNIX Timestamp 2022-06-14 17:44:33    5 secs       
 3     123   345 UNIX Timestamp 2022-06-14 17:44:37    5 secs       
 4     124   346 UNIX Timestamp 2022-06-14 15:50:10 2390 secs       
 5     124   346 UNIX Timestamp 2022-06-14 15:51:01 2390 secs       
 6     124   346 UNIX Timestamp 2022-06-14 16:30:00 2390 secs       
 7     125   345 UNIX Timestamp 2022-06-14 23:55:30  329 secs       
 8     125   345 UNIX Timestamp 2022-06-14 23:58:50  329 secs       
 9     125   345 UNIX Timestamp 2022-06-14 23:59:45  329 secs       
10     125   345 UNIX Timestamp 2022-06-15 00:00:32  329 secs       
11     125   345 UNIX Timestamp 2022-06-15 00:00:59  329 secs       

標籤：r 约会时间 Unix 时间戳 点击流

上一篇：“TypeError:'datetime.datetime'物件不可呼叫”關于使用discord.py制作不和諧機器人

下一篇：熊貓生日不承認六月節