如何找到因子中两个日期之间的差异？

Question

我有一个名为“enrollments”的数据框：

enrolled_at，unenrolled_at和fully_participated_at是因素。我想在我的数据框中添加一个新列，指示两个非空属性之间的小时差异。此新列的类型并不重要，但必须以此格式显示时间（HH MM SS）。

我想做以下伪代码：

If (unenrolled_at == empty && fully_participated_at != empty) 
    newAttributeValue = fully_participated_at - enrolled_at
else if (unenrolled_at != empty && fully_participated_at == empty)
    newAttributeValue = unenrolled_at - enrolled_at
else
    do nothing

编辑：我尝试了网站中的所有方法来执行此操作，但它们不起作用。时间存储在我的数据帧中作为因子类，但站点中的解决方案是因子因子或（字符串）时间 - （字符串）时间。我也分别尝试了“as.character”和“as.Date”函数。所以我的问题不重复。 Rolando Tamayo提供了不同的方法来解决我的问题，但它给了我错误：“ymd_hms中的错误（注释$ unenrolled_at）：找不到函数”ymd_hms“”（我安装了lubridate包）

Answer 1

您可以使用包lubridate：

library(lubridate)


#Create a df with dates

df<-tibble::tibble(
  enrolled_at=as.factor(c("2002-06-09 12:45:40 UTC","2003-01-29 09:30:40 UTC",
                         "2002-09-04 16:45:40 UTC")),
 unenrolled_at=as.factor(c("2002-11-13 20:00:40 UTC",
                        "2002-07-07 17:30:40","2002-07-07 17:30:40 UTC")))
df

# A tibble: 3 x 2
              enrolled_at           unenrolled_at
                   <fctr>                  <fctr>
1 2002-06-09 12:45:40 UTC 2002-11-13 20:00:40 UTC
2 2003-01-29 09:30:40 UTC     2002-07-07 17:30:40
3 2002-09-04 16:45:40 UTC 2002-07-07 17:30:40 UTC

#Check Class
class(df$enrolled_at)

[1] "factor"

#Check class after function ymd_hms
class(ymd_hms(df$enrolled_at))

[1] "POSIXct" "POSIXt"

#Calculete de difference in days
dif<-ymd_hms(df$ unenrolled_at)-ymd_hms(df$enrolled_at)

#difference like a period
as.period(dif)

 [1] "157d 7H 15M 0S"    "-205d -16H 0M 0S"  "-58d -23H -15M 0S"

#Add as a column in df
df$newAttributeValue<-as.period(ymd_hms(df$ unenrolled_at)-ymd_hms(df$enrolled_at))

df

# A tibble: 3 x 3
              enrolled_at           unenrolled_at newAttributeValue
                   <fctr>                  <fctr>      <S4: Period>
1 2002-06-09 12:45:40 UTC 2002-11-13 20:00:40 UTC    157d 7H 15M 0S
2 2003-01-29 09:30:40 UTC     2002-07-07 17:30:40  -205d -16H 0M 0S
3 2002-09-04 16:45:40 UTC 2002-07-07 17:30:40 UTC -58d -23H -15M 0S