R - Time Series Filtration

Question

I have a time series

mainTimeSeries <- data.frame(time=seq(as.POSIXct("2012/1/1"), as.POSIXct("2012/1/5"), "hour"), value=sample(1:10, 1)) 

I want to filter out all the data that is in another series

 badTimeSeries<-data.frame(startTime=seq(as.POSIXct("2012/1/3"), as.POSIXct("2012/1/4"), "hour")) badTimeSeries$endTime <- badTimeSeries$startTime + 1800 

Is there an existing function for filtering dates? The results should be such that no mainTimeSeries element should be between startTime and endTime for badTimeSeries.

Answer 1

lubridate is useful here. Without this, you need to write your own overlap checks, which are pain ...

 library(lubridate) badRange <- as.interval(days(1), as.POSIXct("2012/1/3")) > mainTimeSeries %within% badRange [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE [22] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE [43] FALSE FALSE FALSE FALSE FALSE FALSE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE [64] TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE [85] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE > 

Or:

 > mainTimeSeries[mainTimeSeries %within% badRange] [1] "2012-01-03 00:00:00 PST" "2012-01-03 01:00:00 PST" "2012-01-03 02:00:00 PST" "2012-01-03 03:00:00 PST" [5] "2012-01-03 04:00:00 PST" "2012-01-03 05:00:00 PST" "2012-01-03 06:00:00 PST" "2012-01-03 07:00:00 PST" [9] "2012-01-03 08:00:00 PST" "2012-01-03 09:00:00 PST" "2012-01-03 10:00:00 PST" "2012-01-03 11:00:00 PST" [13] "2012-01-03 12:00:00 PST" "2012-01-03 13:00:00 PST" "2012-01-03 14:00:00 PST" "2012-01-03 15:00:00 PST" [17] "2012-01-03 16:00:00 PST" "2012-01-03 17:00:00 PST" "2012-01-03 18:00:00 PST" "2012-01-03 19:00:00 PST" [21] "2012-01-03 20:00:00 PST" "2012-01-03 21:00:00 PST" "2012-01-03 22:00:00 PST" "2012-01-03 23:00:00 PST" [25] "2012-01-04 00:00:00 PST" > 

---

Using only the R base:

 bad_start <- as.POSIXct('2012/1/3') bad_end <- as.POSIXct('2012/1/4') mainTimeSeries[mainTimeSeries > bad_end | mainTimeSeries < bad_start]