rust - 我如何通过询问 "can it be turned into a number"来移植识别值属于哪个 CSV header 的 Python 代码？

Question

这些值本质上是 CSV 中的堆叠列，它们都出现在同一列中。为了让我辨别一个值属于哪个 header ，我几乎只是说“它可以变成一个数字”。

我在 Python 中尝试做的最简单的例子:

data = ["a","1","2","3","b","4","5","6","c","7","8","9"]

class treatment:
    def __init__(self, name):
        self.name = name
        self.data = []

results = []

for d in data:
    try:
        int(d)
    except ValueError:
        results.append(treatment(name=d))
    else:
        results[-1].data.append(d)

这是我在 Rust 中拥有的，但是在解析后不起作用:

extern crate csv;

use std::error::Error;
use std::fs::File;
use std::io;
use std::prelude::v1::Option;
use std::process;
use std::vec::Vec;

struct Treatment {
    raw_name: String,
    channel: String,
    data: Vec<i64>,
}

impl Treatment {
    fn observe(&mut self, observation: i64) {
        &mut self.data.push(observation);
    }
}

struct Experiment {
    treatments: Vec<Treatment>,
}

impl Experiment {
    fn add_treatment(&mut self, treatment_name: &str, channel: &str) {
        &mut self.treatments.push(Treatment {
            raw_name: treatment_name.to_string(),
            channel: channel.to_string(),
            data: Vec::new(),
        });
    }

    fn get_current_experiment(&mut self) -> &Treatment {
        self.treatments.last().unwrap()
    }
}

fn deserialize_csv(file_path: String) -> csv::Reader<File> {
    csv::ReaderBuilder::new()
        .has_headers(false)
        .from_path(file_path)
        .expect("Experienced issues reading csv file")
}

fn vectorize_string_records(mut csv_reader: csv::Reader<File>) {
    let mut experiments: Experiment = Experiment {
        treatments: Vec::new(),
    };

    for row in csv_reader.records() {
        // for these datasets we only care about data in the first cell
        let datapoint = match &row {
            Ok(T) => T.get(0),
            Err(E) => Some(""),
        }.unwrap();

        let val_as_int = datapoint.parse::<i64>();

        let current_treatment = experiments.get_current_experiment();

        // discern if value is header or datapoint
        match val_as_int {
            Ok(int_val) => &current_treatment.observe(int_val),
            Err(new_header) => &experiments.add_treatment(datapoint, &"ch=1".to_string()),
        };

        println!("{}", datapoint);
    }
}

fn main() {
    let i = deserialize_csv("/home/file.csv".to_string());
    vectorize_string_records(i);
}

在尝试用 Rust 实现它时，我在一些概念上遇到了困难。

Answer 1

将您的 Python 代码直接移植到 Rust:

#[derive(Debug)]
struct Treatment {
    name: String,
    data: Vec<i32>,
}

impl Treatment {
    fn new(name: String) -> Self {
        Self {
            name,
            data: Vec::new(),
        }
    }
}

fn main() {
    let data = ["a", "1", "2", "3", "b", "4", "5", "6", "c", "7", "8", "9"];
    let mut results: Vec<Treatment> = Vec::new();
    for d in &data {
        if let Ok(n) = d.parse::<i32>() {
            let last = results
                .last_mut()
                .expect("Some header must come before the values");
            last.data.push(n);
        } else {
            results.push(Treatment::new(d.to_string()))
        }
    }
    println!("results = {:#?}", results);
}

Playground