```{r setup}
library(tidyverse)
library(RcppRoll)
```


```{r function}

splatter <- function(var, fun, n = 10, ...) {

  args <- quos(...)
  var  <- enquo(var)
  fun  <- enquo(fun)


  if(length(n) == 1) {
    i <- seq_len(n)
  } else {
    i <- n
  }

  map( i, ~quo(UQ(fun)(!!var, !!.x, !!!args)) ) %>%
    set_names(sprintf("%s_%s_%s", quo_text(var), quo_text(fun), i))

}

```


```{r output}

library(gapminder)

rolls <- splatter(lifeExp, roll_mean, c(2, 3, 5), fill = NA)


gapminder %>% group_by(country) %>% mutate(!!!rolls) %>%
  select(country, year, matches("roll_mean"))

# A tibble: 1,704 x 5
# Groups:   country [142]
   country      year lifeExp_roll_mean_2 lifeExp_roll_mean_3 lifeExp_roll_mean_5
   <fct>       <int>               <dbl>               <dbl>               <dbl>
 1 Afghanistan  1952                29.6                NA                  NA
 2 Afghanistan  1957                31.2                30.4                NA
 3 Afghanistan  1962                33.0                32.1                32.2
 4 Afghanistan  1967                35.1                34.0                34.2
 5 Afghanistan  1972                37.3                36.2                36.1
 6 Afghanistan  1977                39.1                38.1                37.8
 7 Afghanistan  1982                40.3                39.7                39.4
 8 Afghanistan  1987                41.2                40.8                40.5
 9 Afghanistan  1992                41.7                41.4                41.2
10 Afghanistan  1997                41.9                41.9                42.0
# ... with 1,694 more rows

```