Hi, I am working on github issue #2074 as a first time contributor and a beginner to Rust. I am difficulties to get the ops/utf8.rs method compiled. Can someone help me out with this?

Hi! What issues are you having? Would you like to post the errors you’re having?

Hi, I am trying to implement the functionality like this in path Daftsrcdaft-coresrcarrayopsutf8.rs

pub fn repeat(&self, n: usize) -> DaftResult {
    let self_arrow = self.as_arrow();

    // Handle empty data case.
    if self.is_empty() {
        return Ok(Utf8Array::empty(self.name(), &DataType::Utf8));
    }
    let arrow_result = self_arrow
        .iter()
        .flat_map(|element| element.map(|w| w.repeat(n)))
        .collect::>>();
    
    Ok(Utf8Array::from((self.name(), Box::new(arrow_result?))))
}

For which I get the following Error output:

rror[E0277]: a value of type `Result<arrow2::array::Utf8Array<i64>, DaftError>` cannot be built from an iterator over elements of type `std::string::String`
   --> src\daft-core\src\array\ops\utf8.rs:510:14
    |
510 |             .collect::<DaftResult<arrow2::array::Utf8Array<i64>>>();
    |              ^^^^^^^ value of type `Result<arrow2::array::Utf8Array<i64>, DaftError>` cannot be built from `std::iter::Iterator<Item=std::string::String>`
    |
    = help: the trait `FromIterator<std::string::String>` is not implemented for `Result<arrow2::array::Utf8Array<i64>, DaftError>`
    = help: the trait `FromIterator<Result<A, E>>` is implemented for `Result<V, E>`

Hey @Akshay Verma! Instead of

flat_map

, try using

map

instead, i.e.

let arrow_result = self_arrow
            .iter()
            .map(|element| element.map(|w| w.repeat(n)))
            .collect::<arrow2::array::Utf8Array<i64>>();

what's happening with your code is that

flat_map

is filtering out the null values, resulting in an iterator of

Strings

. This won't work because arrow2 arrays are built from

Option<String>

, and also because we want to preserve cardinality in the result array.

Thanks. That makes sense. I am trying out the build. currently getting errors in openssl build (I am working on a Windows system). Let me get back on it.

Hi @Akshay Verma! You might be hitting a merge conflict for a recent change in our

to_field

function signature. You can reference the other utf8

to_field

implementations, but those inputs are now

ExprRef

!

Let me try that

It seems that you should change your

utf8_repeat

function to take a Series instead of a single number You can reference the other utf8 kernels for an example. A good one to check out might be something like `left`:

Daft/src/daft-dsl/src/functions/utf8/left.rs

Thanks, I was looking at right.rs 🙂

Yup! The function signature of right (in

Daft/src/daft-core/src/series/ops/utf8.rs

):

pub fn utf8_right(&self, nchars: &Series) -> DaftResult<Series>