【问题标题】:How to add millisecond on timestamp correctly using R如何使用 R 在时间戳上正确添加毫秒
【发布时间】:2014-08-29 08:40:49
【问题描述】:

tick_count 列的单位是 100 毫秒。我想在 date_time 上加上 tick_count 时间。但结果没有正确执行加法。 例如,第二个值应该是“2014-08-15 12:51:02.6 EEST”而不是“2014-08-15 12:51:02.5 EEST”,我不知道为什么会这样。它对我的结果造成了严重影响,因为有时它会导致两三行有相同的时间。 这种情况该如何处理?

>head(status306$date_time)
#[1] "2014-08-15 12:51:01 EEST" "2014-08-15 12:51:01 EEST" "2014-08-15 12:51:01 EEST"
#[4] "2014-08-15 12:51:01 EEST" "2014-08-15 12:51:01 EEST" "2014-08-15 12:51:01 EEST"
> head(status306$tick_count/10)
#[1] 0.0 1.6 1.7 1.1 2.6 2.8
> head(status306$date_time +status306$tick_count/10)
#[1] "2014-08-15 12:51:01.0 EEST" "2014-08-15 12:51:02.5 EEST" "2014-08-15 12:51:02.7 EEST"
#[4] "2014-08-15 12:51:02.0 EEST" "2014-08-15 12:51:03.5 EEST" "2014-08-15 12:51:03.7 EEST"

更新

这是原始数据框,我才知道:p

structure(list(grint = c("C51CCF4CAACGGCC44", "C55CCF4CCCCGGCC44", 
"C55CCF4CCCCGGCC44", "C55CCF4CCCCGGCC44", "C55CCF4CCC011CC44", 
"C55CCF4CCC011CC44"), dint = c("0000010100001010000000000001110000000001100001000000", 
"0000010100001010000000000001110000000001100001000000", "0100010100001010000000000001110000000001100001000000", 
"0100010100001011000000000001110000000001100001000000", "0000010100001011000000000001110000000001100001000000", 
"0000010100001011000000000000100000000001100001000000"), tick_count = c(0L, 
16L, 17L, 11L, 26L, 28L), date_time = structure(c(1408096261, 
1408096261, 1408096261, 1408096261, 1408096261, 1408096261), class = c("POSIXct", 
"POSIXt"), tzone = ""), detector_signal_levels = c("135,1,163,1,165,5,167,167,5,167,5,163,3,169,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,", 
"135,5,163,1,165,5,167,167,5,167,5,163,3,169,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,", 
"135,5,163,1,165,5,167,167,5,167,5,163,3,169,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,", 
"135,5,163,1,165,5,167,7,5,167,5,163,3,169,77,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,", 
"135,5,163,1,165,5,167,7,9,167,1,163,3,169,121,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,", 
"135,5,163,1,165,5,167,7,9,15,1,163,3,169,123,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,"
)), .Names = c("grint", "dint", "tick_count", "date_time", "detector_signal_levels"
), row.names = c(NA, 6L), class = "data.frame")

【问题讨论】:

  • 我怀疑原因是您的 $date_time 值包括毫秒,因此您没有添加您认为添加的内容。显示的内容不一定与实际值具有相同的精度。 $date_time 是 POSIX 级别的吗?
  • 是的,我认为它是 POSIX 类。这很奇怪,但我不认为原因是 $date_time 值包含毫秒,因为实际上对于序列数据,它们是在一个数据包中发送的,date_time 是接收数据包的时间,时间差大约是 $tick_count .
  • @Chenlu 我用你的dput数据集更新了

标签: r timestamp posixct


【解决方案1】:

试试:(如果你的输入是characterclass)

date_time <- c("2014-08-15 12:51:01", "2014-08-15 12:51:01", "2014-08-15 12:51:01",
         "2014-08-15 12:51:01", "2014-08-15 12:51:01", "2014-08-15 12:51:01")

date_time1 <- strptime(date_time, format="%Y-%m-%d %H:%M:%S", tz="EEST")
v1 <- c(0.0, 1.6,1.7,1.1, 2.6, 2.8)

date_time1$sec <- date_time1$sec+v1
op <- options(digits.secs=3)
date_time1
#[1] "2014-08-15 12:51:01.0 EEST" "2014-08-15 12:51:02.6 EEST"
#[3] "2014-08-15 12:51:02.7 EEST" "2014-08-15 12:51:02.1 EEST"
#[5] "2014-08-15 12:51:03.6 EEST" "2014-08-15 12:51:03.8 EEST"

我想我可以重现你的结果:

date_time1 <- as.POSIXct(date_time, format="%Y-%m-%d %H:%M:%S", tz="EEST")
date_time1+v1
#[1] "2014-08-15 12:51:01.0 EEST" "2014-08-15 12:51:02.5 EEST"
#[3] "2014-08-15 12:51:02.7 EEST" "2014-08-15 12:51:02.0 EEST"
#[5] "2014-08-15 12:51:03.5 EEST" "2014-08-15 12:51:03.7 EEST"

更新

转换成POSIXlt就可以得到你预期的结果(不过我不确定你预期的结果是否正确)

 status306$date_time<- as.POSIXlt(status306$date_time)
 status306$date_time$sec <- with(status306, date_time$sec + tick_count/10)
 status306$date_time
 #[1] "2014-08-15 05:51:01.0 EDT" "2014-08-15 05:51:02.6 EDT"
 #[3] "2014-08-15 05:51:02.7 EDT" "2014-08-15 05:51:02.1 EDT"
 #[5] "2014-08-15 05:51:03.6 EDT" "2014-08-15 05:51:03.8 EDT"

更新2

  status306$date_time[4] <- as.POSIXct("2014-08-15 12:53:59")
  status306$date_time<- as.POSIXlt(status306$date_time)
  val <- with(status306, date_time$sec + tick_count/10)
  status306$date_time$sec <- ifelse(val >=60, val-60, val)
  status306$date_time$min <- ifelse(val >=60, status306$date_time$min+1, status306$date_time$min)

  status306$date_time
  #[1] "2014-08-15 05:51:01.0 EDT" "2014-08-15 05:51:02.6 EDT"
  #[3] "2014-08-15 05:51:02.7 EDT" "2014-08-15 12:54:00.1 EDT"
  #[5] "2014-08-15 05:51:03.6 EDT" "2014-08-15 05:51:03.8 EDT"



 options(op)

【讨论】:

  • 这并不能解决问题,因为您已经创建了字符串作为输入。
  • @Carl Witthoft 用户没有使用dput。所以,这对我来说是唯一的选择,由于声誉,我无法发表评论。此外,字符串由strptime 转换。从OP的帖子中不清楚是POSIXct还是POSIXlt
  • 没关系,但您应该在答案中添加一些文本,例如“如果您的输入是字符,那么...”
  • @Carl Witthoft,对不起,我没有说,输入的 date_time 不是字符串,它是 POSIXct,类型是 double。顺便说一句,什么是“dput”,如果方便的话,我想了解更多
  • @Chenlu 您可以使用dput(head(status306)) 并将输出复制并粘贴到您的帖子中,以便其他人可以重现您原始数据集的结构。
猜你喜欢
  • 1970-01-01
  • 2020-12-16
  • 2013-06-22
  • 1970-01-01
  • 2022-10-25
  • 1970-01-01
  • 1970-01-01
  • 2019-08-01
相关资源
最近更新 更多