34  Vectorize gh_neighbor()

Objective

Vectorize the gh_neighbor() function and make it robust to errors.

We’ll now focus on vectorizing the gh_neighbor() function using the same approach we took for gh_encode().

Exercise 1

The objective of this exercise is to vectorize the gh_neighbor() function. You an do this by:

• Change the argument geohash to take a character vector via geohash: Strings
• Return a character vector using -> Strings
• Add the function to extendr_module! { }
• Test the function on new data in R

View solution

#[extendr]
fn gh_neighbor(geohash: Strings, direction: String) -> Strings {
    let dir = as_direction(direction);
    geohash
        .into_iter()
        .map(|gh| neighbor(&gh, dir).unwrap())
        .collect::<Strings>()
}

In R test your function with:

n <- 1000
x <- runif(n, -180, 180)
y <- runif(n, -90, 90)

geohashes <- gh_encode(x, y, 8)

gh_neighbor(geohashes, "e")

Exercise 2

Make this function robust to errors by returning an NA when an error is encountered.

View solution

#[extendr]
fn gh_neighbor(geohash: Strings, direction: String) -> Strings {
    let dir = as_direction(direction);
    geohash
        .into_iter()
        .map(|gh| {
            let nb = neighbor(&gh, dir);
            match nb {
                Ok(res) => Rstr::from(res),
                Err(_) => Rstr::na(),
            }
        })
        .collect::<Strings>()
}