I often like to get a feel for a text file containing a lot of numbers quickly by making a histogram. One can do this in plotting software like Gnuplot, but sometimes having something in the terminal is nice (or can even make plotting later easier).

To this end, I wrote a program that reads numbers from stdin and sorts them into bins of equal width. The output of the program is the bin centers and counts in each bin.

I am new to Rust. I found the clap library for parsing command line arguments, and I tried to use some language features for manipulating vectors.

I suspect the algorithm can be improved. Because the code outputs the bins sorted by bin center, I store the bin edges in a vector as they are created. However, to avoid comparing floats when checking if a bin is already present, I have another parallel vector with bin ID numbers that I compare instead.

// binner
//  Command-line tool for reading in a list of numbers and writing 
//  bins and counts for a specified bin width 
//  bins are defined [lower, upper), such that each output point
//  has the form:
//  (x,y) = ((lower + upper) / 2, counts)
//  and lower - upper = specified binwidth

extern crate clap;

use std::io::{self, BufRead};
use std::cmp::Ordering;

use clap::{Arg, App}; // for parsing command-line arguments

fn bins(values: &Vec<f64>, binwidth: f64, binstart: f64) -> Vec<(f64, u32)> {
    // Input: a vector of floats and a bin width 
    // Output: a vector of bin edges and counts for each bin

    // unique integer identifier for each bin, to be used to check if bin is present
    // the actual bin edges will be computed as floats, so this avoids 
    // checking for strict float equality
    let mut bin_ids: Vec<i32> = Vec::new();

    // computed bin edge values and counts 
    let mut edges_counts: Vec<(f64, u32)> = Vec::new();

    for val in values {
        let bin_id = (val / binwidth).floor() as i32;

        // check if the bin is already in the vector
        match bin_ids.iter().position(|&b| b == bin_id) {
            // if it is, add to the corresponding counts
            Some(i) => { edges_counts[i].1 += 1 },
            // else, add a new bin
            _ => {
                    (binwidth * (((val - binstart) / 
                      binwidth).floor() + 0.5) + binstart, 1)

    // sort result by edge, then return. Since floats may contain NaN,
    // we need to use partial_cmp since edges are floats,
    // and specify what to do for errors
    edges_counts.sort_by(|a, b| a.0.partial_cmp(&b.0).unwrap_or(Ordering::Equal));

    return edges_counts;

fn main() {
    // parse command line arguments and set up help with clap-rs
    let arg_matches = App::new("binner")
        .about("Read numbers from standard input, output bins and counts")
             .help("Set bin width")
             .value_name("EDGE START")
             .help("Set bin edge start")
             .help("Input stream")

    // type-checking macro explaind in clap example 12_typed_values.rs
    let binwidth: f64 = value_t!(arg_matches, "binwidth", f64).unwrap_or(1.0);
    let binstart: f64 = value_t!(arg_matches, "binstart", f64).unwrap_or(0.0);

    // parse stdin to make a list of values 
    let stdin = io::stdin();
    let mut values: Vec<f64> = Vec::new();
    for line in stdin.lock().lines() {
        let elem: f64 = match line.unwrap().trim().parse() {
            Ok(num) => num,
            Err(_) => {
                eprintln!("Invalid value entered");

    // compute bins and print
    let result = bins(&values, binwidth, binstart); 
    for bin in &result {
        println!("{}\t{}", bin.0, bin.1);

mod tests{
    use super::bins;

    fn sequence() {
        // bin 10 numbers. bin edges start at binwdith / 2 intervals
        let vals = vec![1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5];
        let out = vec![(1.5, 2), (2.5, 2), (3.5, 2), (4.5, 2), (5.5, 2)];
        assert_eq!(bins(&vals, 1.0, 1.0), out); 

    fn unordered() {
        let vals = vec![1.0, 55.6, -15.2, 55.9];
        let out = vec![(-15.5, 1), (1.5, 1), (55.5, 2)];
        assert_eq!(bins(&vals, 1.0, 1.0), out); 

A simple use-case might be

seq 0 1 1000 | binner -w 10

which bins integers from 0 to 1000 in bins of width 10.


1 Answer 1



//  and lower - upper = specified binwidth

Wait a minute! Strike that, reverse it. Thank you.

Also typo nit: "explaind"


Please introduce a helper that maps from val to bin_id. I am skeptical that bin ids are compatible with edge counts currently, since binstart does not participate in computing a bin_id. This matters when binstart is not a multiple of binwidth.


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.