feat: Implement Spark function space

[kazantsev-maksim]

Which issue does this PR close?

• N/A

Rationale for this change

Add new function: https://spark.apache.org/docs/latest/api/sql/index.html#space

What changes are included in this PR?

• Implementation
• Unit Tests
• SLT tests

Are these changes tested?

Yes, tests added as part of this PR.

Are there any user-facing changes?

No, these are new function.

[andygrove]

I think that space is quite often invoked with a literal argument, e.g., space(2). The current implementation converts the scalar value into an array for each batch and then calls " ".repeat(m as usize) for each row, resulting in the same string being allocated each time. This could be optimized in the scalar case to build the string once rather than per row and then append the same string to the builder repeatedly.

[andygrove]

In fact, the return type in this case could be a scalar not an array

[andygrove]

This allocates a new string per row, which is discarded after being appended to the array. It would be more efficient for this function to build the string buffer and offsets directly.

[andygrove]

Could you add tests that use array inputs. The tests currently only cover scalar inputs.

[kazantsev-maksim]

@andygrove thanks for the review, the benchmark test result in Comet is still not ideal.

[comphead]

Thanks @kazantsev-maksim just thinking aloud if we can iterate values only once?

fn spark_space_array_inner(array: &Int32Array) -> StringArray {
    let mut builder = StringBuilder::new(array.len());
    let mut space_buf = String::new();

    for v in array.iter() {
        match v {
            None => builder.append_null(),
            Some(l) if *l > 0 => {
                let l = *l as usize;
                if space_buf.len() < l {
                    space_buf = " ".repeat(l);
                }
                builder.append_value(&space_buf[..l]);
            }
            Some(_) => builder.append_value(""),
        }
    }

    builder.finish()
}

something like that?

[kazantsev-maksim]

@comphead thanks, performance has improved!

[comphead]

Thanks @kazantsev-maksim and @andygrove for the review