[R] How to calculate running 8-hour averages of temperature

[arun]

Hi,
You may try
?rollmean from library(zoo)
?SMA from library(TTR)
dat1 <- structure(...
library(zoo)

res1 <- rollmean(dat1[-c(1:5),2],8)
library(TTR)
res2 <- SMA(dat1[-c(1:5),2],8)
 all.equal(res1,res2[!is.na(res2)])
#[1] TRUE



A.K.

Hi, 

How can I calculate running 8-hour averages of temperature 
values that start at specific date and time? In the example data set 
shown below I want to start the calculation on May 31, at 5pm (row 6). 
Reproducible data is provided. 

Thanks 
structure(list(date = structure(c(644148000, 644151600, 644155200, 
644158800, 644162400, 644166000, 644169600, 644173200, 644176800, 
644180400, 644184000, 644187600, 644191200, 644194800, 644198400, 
644202000, 644205600, 644209200, 644212800, 644216400, 644220000, 
644223600, 644227200, 644230800, 644234400, 644238000, 644241600, 
644245200, 644248800, 644252400, 644256000), class = c("POSIXct", 
"POSIXt"), tzone = ""), temp = c(17.7, 18.3, 18.9, 19.7, 19.5, 
19, 18.8, 18, 16.3, 15.1, 13.9, 12.7, 11.9, 10.6, 10, 8.7, 7.89999, 
8.09999, 9.29999, 12.4, 14.5, 15.9, 17.4, 18.6, 18.3, 18.4, 18.5, 
20.4, 19.7, 18.9, 18.1)), .Names = c("date", "temp"), row.names = 3613:3643, class = "data.frame")