4
\$\begingroup\$

In Rust, I want to take an array of u32 values, convert each to four bytes in big endian, and concatenate them to yield a Vec<u8> result. Example:

let input: [u32; 5] = [
    0x67452301, 0xEFCDAB89, 0x98BADCFE, 0x10325476, 0xC3D2E1F0];

(... Do some data conversion work ...)

let output: Vec<u8> = vec![
    0x67, 0x45, 0x23, 0x01,
    0xEF, 0xCD, 0xAB, 0x89,
    0x98, 0xBA, 0xDC, 0xFE,
    0x10, 0x32, 0x54, 0x76,
    0xC3, 0xD2, 0xE1, 0xF0,
];

I have several pieces of working code that all give the same answer.

Algorithm A (for-loop):

let mut output = Vec::<u8>::new();
for val in &input{
    output.extend_from_slice(&val.to_be_bytes());
}

Algorithm B (for-each):

let mut output = Vec::<u8>::new();
input.iter().for_each(|val| output.extend_from_slice(&val.to_be_bytes()));

Algorithm C (fold-append):

let output = input.iter().fold(Vec::<u8>::new(),
    |mut acc, val| {acc.extend_from_slice(&val.to_be_bytes()); acc});

Algorithm D (flat-map):

let output: Vec<u8> = input.iter().flat_map(|val| val.to_be_bytes().to_vec()).collect();

Is there an even shorter, clearer way to accomplish the task? Maybe algorithm D's to_vec() could be avoided somehow?

Speed is not a concern because this code only executes a few times, and the input array length is fixed at 5.

\$\endgroup\$

2 Answers 2

3
\$\begingroup\$

Algorithm A seems the clearest to me. An improvement might be to use Vec::with_capacity to avoid the allocation. Indeed, arrays in Rust are currently somewhat cumbersome to use.

My advice would be to wrap it in a function and not worry about it later on:

pub fn to_bytes(input: &[u32]) -> Vec<u8> {
    let mut bytes = Vec::with_capacity(4 * input.len());

    for value in input {
        bytes.extend(&value.to_be_bytes());
    }

    bytes
}

#[cfg(test)]
mod tests {
    #[test]
    fn to_bytes() {
        use super::*;

        assert_eq!(to_bytes(&[]), &[]);

        let input = &[0x67452301, 0xEFCDAB89, 0x98BADCFE, 0x10325476, 0xC3D2E1F0];
        let output = &[
            0x67, 0x45, 0x23, 0x01, 0xEF, 0xCD, 0xAB, 0x89, 0x98, 0xBA, 0xDC, 0xFE, 0x10, 0x32,
            0x54, 0x76, 0xC3, 0xD2, 0xE1, 0xF0,
        ];

        assert_eq!(to_bytes(input), output);
    }
}

(playground)

\$\endgroup\$
3
\$\begingroup\$

Rust 1.53.0 (2021-06-17) introduces the IntoIterator for array types, which finally makes this shorter code possible:

Algorithm E (flat-map simpler):

let output: Vec<u8> = input.iter().flat_map(|val| val.to_be_bytes()).collect();

(This is based on my algorithm D, removing .to_vec(). I did check that algorithm E fails to compile using Rust 1.52.0.)


Algorithm F (not recommended, Rust 1.51.0, see std::array::IntoIter::new()):

let output: Vec<u8> = input.iter().flat_map(|val|
    std::array::IntoIter::new(val.to_be_bytes())).collect();
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.