29  Vectorization

TipObjective

Take a character vector and create a list of external pointers.

In the previous exercise we worked with a single url. In R we typically work with a vector of values. Naturally, we won’t want to work with just a single url at a time, we want to work with hundreds and thousands of them.

So, how can we handle this? Through vectorization (a fancy name for iteration and looping).

Now, an external pointer isn’t any type of vector that im aware of! So how do we store them?

Remember!! Lists are vectors (see this video).

Lists can hold any R object in them that they want!! So this is a natural way to contain multiple external pointers at a time!!!

What we can do is use the Strings wrapper type (think of this as an NA-aware version of Vec<String>).

Strings wrapper type

We can use .into_iter() to iterate through it. Each element is an Rstr, which is an NA aware &str—it is able to be used in most places that &str can be used.

Collecting into a list

  • We can .collect() directly into a List
  • Every single wrapper type can be turned into an Robj (because everything is an SEXP!)
  • Our external pointer can be turned into an Robj

For example:

#[extendr]
fn parse_shapes(shapes: Strings) -> List {
    shapes
        .into_iter()
        .map(|vi| RShape::new(vi).unwrap())
        .collect()
}

Exercise

Take a character vector (Strings) and create a list of external pointers.

  • Use .unwrap() to ignore the result for now
  • Use .collect::<List>() to capture the results into a List
View solution 
#[extendr]
fn url_parse(url: Strings) -> List {
    url.into_iter().map(|xi| {
        RUrl::new(xi).unwrap()
    }).collect::<List>()
}