Improving for loop inception speed when comparing Date-time values - r

Question

I am looking to speed up some code of mine, and am looking for some advice. I have two dataframes, and want to take data from one and input it into the other, but their formats are quite different. I have written some code that works, but it runs fairly slow, especially when I have large datasets. Example dataframes are below:
dat = structure(list(Date = structure(c(1508641200, 1508643000, 1508644800,
1508646600, 1508648400, 1508650200, 1508652000, 1508653800, 1508655600,
1508657400, 1508659200, 1508661000, 1508662800, 1508664600, 1508666400,
1508668200, 1508670000, 1508671800, 1508673600, 1508675400, 1508677200,
1508679000, 1508680800, 1508682600, 1508684400, 1508686200, 1508688000,
1508689800, 1508691600, 1508693400), class = c("POSIXct", "POSIXt"
), tzone = "EST"), X = c(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), tru = c(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("Date", "X", "tru"), row.names = c(NA,
30L), class = "data.frame")
and,
truth = structure(list(startdate = structure(c(1509937620, 1510705200,
1510722240, 1512245160, 1512250560, 1512251760, 1512271140, 1512274440,
1512984360, 1512986760, 1513002600, 1513004700, 1513752000, 1513753800,
1513775940, 1513776840, 1514268900, 1514289000, 1514307900, 1517842620,
1518056460, 1520031660, 1520050560, 1520075580, 1520218620), class =c("POSIXct",
"POSIXt"), tzone = ""), enddate = structure(c(1509952320, 1510718040,
1510737240, 1512250260, 1512269640, 1512269940, 1512290280, 1512289380,
1513001100, 1513001400, 1513023840, 1513021440, 1513772640, 1513771440,
1513797180, 1513796280, 1514288400, 1514307600, 1514311140, 1517857320,
1518067560, 1520049060, 1520073240, 1520078880, 1520222820), class =c("POSIXct",
"POSIXt"), tzone = ""), Y = c(42340.1667145732, 49765.2381579195,
48687.3848496384, 31170.9693454847, 50435.3541955455, 49757.5112973802,
44031.8550803252, 45912.1378875664, 47193.1529894274, 49826.4304479959,
45840.7120690243, 42483.44259103, 52188.4048476908, 52783.8164119854,
51769.0550080142, 49866.1301140174, 52250.5531316799, 49754.6933212176,
45256.185763228, 47742.0544890968, 53414.0366523465, 51881.9495162963,
48632.3656223053, 44476.3677890439, 61922.0098972212)), .Names =
c("startdate",
"enddate", "Y"), row.names = c(NA, 25L), class = "data.frame")
Now, what I want to do is take values from truth$Y and input them into dat$tru at their appropriate date-time values given as these actions happen between truth$startdate and truth$enddate.
What I have currently is the following code to do so:
for(i in 1:length(truth$startdate)){
for(j in 1:length(dat$Date)){
if(dat$Date[j] >= truth$startdate[i] && dat$Date[j] <= truth$enddate[i]){
dat$tru[j] = truth$Y[i]
}
}
}
Is there a more efficient way to do this avoiding a for loop inside of a for loop? Thanks in advance!

Answer 1

Using the between function from the dplyr package, you can eliminate the inner loop:
library(dplyr)
for(i in 1:length(truth$startdate)){
dat$tru[between(dat$Date, truth$startdate[i], truth$enddate[i])] <-truth$Y[i]
}
This should result in a significant performance improvement.
I sure this could be further refined but this is a quick and easy fix. I suggest looking at the purrr package, it could offer some additional performance improvements.