Let's say I have hourly (6 hours in total) wave height data like below. I am aggregating them into 2-hour and 3-hour heights using the mean function
h_1<-c(2, 4, 6, 7, 9, 11)
Here is the function I used:
agg_direct<-function(x){
y<-as.data.frame(x)
names(y)<-c("wh")
agg<-lapply(c(1, 2, 3), function(x)
tapply(y$wh, as.integer(gl(nrow(y), x, nrow(y)) ), FUN = mean))
}
out<-agg_direct(h_1)
Outputs of this:
$h_1
1 2 3 4 5 6
2 4 6 7 9 11
$h_2
1 2 3
3.0 6.5 10.0
$h_3
1 2
4 9
But here is what I exactly want:
For 2-h, mean of (2,4), (4,6), (6,7), (7,9), (9,11)
For 3-h, mean of (2,4,6), (4,6,7), (6,7,9), (7,9,11)
so the outputs will be like:
For 2-h: 3, 5, 6.5, 8, 10
For 3-h: 4, 5.66, 7.33, 9
And this process will be performed following 4-hour, 5-hour...
We can use rollmean from zoo
library(zoo)
rollmean(h_1, 2)
#[1] 3.0 5.0 6.5 8.0 10.0
rollmean(h_1, 3)
#[1] 4.000000 5.666667 7.333333 9.000000