rollmean with dplyr and magrittr

Question

Given the following data:

    set.seed(1)
    data <- data.frame(o=c('a','a','a','a','b','b','b','b','c','c','c','c'), t=c(1,2,3,4,1,2,3,4,1,2,3,4), u=runif(12), v=runif(12))
    data
       o t          u         v
    1  a 1 0.26550866 0.6870228
    2  a 2 0.37212390 0.3841037
    3  a 3 0.57285336 0.7698414
    4  a 4 0.90820779 0.4976992
    5  b 1 0.20168193 0.7176185
    6  b 2 0.89838968 0.9919061
    7  b 3 0.94467527 0.3800352
    8  b 4 0.66079779 0.7774452
    9  c 1 0.62911404 0.9347052
    10 c 2 0.06178627 0.2121425
    11 c 3 0.20597457 0.6516738
    12 c 4 0.17655675 0.1255551

I want to calculate the rolling mean (package zoo) of u per group defined by the coloumn o. The order for the rolling mean is set by t. The rolling mean should be added as a new column to the data.frame.

I want to use magrittr and dplyr. I tried

    data %>%
      group_by(o) %>%
      sort(t) %>%
      select(u) %>%
      rollmean(3) %>%
      rbind

But this won't work. Is it possible to do it with magrittr and dplyr or do I have to do it step by step? The values of o and t are variable in my real data.

How do I fill the first two rows?

Answer 1

May be this helps:

library(dplyr)
data %>%
group_by(o) %>%
mutate(rM=rollmean(u,3, na.pad=TRUE, align="right"))

If you want to do for both columns, u and v

fun1 <- function(x) rollmean(x, 3, na.pad=TRUE, align="right")
data %>% 
group_by(o) %>% 
mutate_each(funs(fun1), u, v)

Answer 2

A more flexible wrapper comes from the rowr package. This allows for windows of different size within your initial data.

data %>% 
group_by(o) %>% 
mutate(MEANS = rollApply(u, fun=mean, window=3, align='right'))