PURUDA

PUre RUst DAtaframe

Compile-time type-safe DataFrame library for Rust. Provides Col1–Col32 generic structs with zero-cost column access, numeric statistics, filtering, sorting, and CSV/JSON I/O.

Quick Start

Add to your Cargo.toml:

[dependencies]
puruda = "0.2"

Example

use puruda::*;

fn main() {
    let x = vec![1, 2, 3];
    let y = vec![4.0f64, 5.0, 6.0];
    let z: Vec<String> = vec!["a", "b", "c"]
        .into_iter().map(|s| s.to_string()).collect();

    let mut df = Col3::from_cols(x, y, z);
    df.set_header(vec!["x", "y", "z"]);

    // Pretty-print table
    println!("{}", df);
    //  x  y  z
    //  -  -  -
    //  1  4  a
    //  2  5  b
    //  3  6  c

    // Shape
    println!("{:?}", df.shape()); // (3, 3)

    // Numeric statistics
    println!("sum: {}", df.c1().sum());       // 6
    println!("mean: {}", df.c2().mean());     // 5.0
    println!("std: {}", df.c2().std_dev());   // 0.8165...

    // Slicing
    let top2 = df.head(2);
    let bottom2 = df.tail(2);

    // Filter rows
    let filtered = df.filter(|i| *df.c1().idx(i) > 1);

    // Sort by column
    let sorted = df.sort_by_c1();

    // Add a row
    df.push_row(4, 7.0, "d".to_string());

    // Transform a column in-place
    df.c2_mut().apply(|v| *v *= 2.0);

    // CSV I/O
    df.write_csv("data.csv", ',').unwrap();
    let df2 = Col3::<Vec<i32>, Vec<f64>, Vec<String>>::read_csv("data.csv", ',').unwrap();

    // JSON I/O
    df.write_json("data.json").unwrap();
    let df3 = Col3::<Vec<i32>, Vec<f64>, Vec<String>>::read_json("data.json").unwrap();
}

API Overview

Column Access & Shape

Method	Description
`c1()` .. `cN()`	Immutable reference to column
`c1_mut()` .. `cN_mut()`	Mutable reference to column
`nrows()`	Number of rows
`ncols()`	Number of columns
`shape()`	`(nrows, ncols)` tuple
`len()`	Alias for `nrows()`
`is_empty()`	`true` if no rows

Slicing

Method	Description
`head(n)`	First `n` rows
`tail(n)`	Last `n` rows
`slice(start, end)`	Rows in `start..end`

Row Operations

Method	Description
`push_row(v1, .., vN)`	Append a single row
`filter(predicate)`	Keep rows where `predicate(row_index)` is true
`append(&other)`	In-place vertical concatenation
`concat(other)`	Returns a new concatenated DataFrame
`reindex(&[usize])`	Rearrange rows by index array

Sorting

Method	Description
`sort_by_c1()` .. `sort_by_cN()`	Sort all rows by column (ascending, requires `Ord`)

Column-Level Traits

Numeric (for Vec<i32>, Vec<f64>, etc.)

Method	Description
`sum()`	Sum of elements
`mean()`	Arithmetic mean (`f64`)
`min_val()`	Minimum value reference
`max_val()`	Maximum value reference
`var()`	Population variance (`f64`)
`std_dev()`	Population standard deviation (`f64`)

ColumnUnique (for types implementing Eq + Hash)

Method	Description
`unique()`	Deduplicated values (preserving order)
`n_unique()`	Count of unique values

ColumnApply

Method	Description
`apply(f)`	Mutate each element in-place

ColumnDisplay

Method	Description
`print()`	Print column as `[v1, v2, ...]`

Utility

Function	Description
`map_column(col, f)`	Map column to a new `Vec<U>` (type conversion)
`describe()`	Print summary table (name, count per column)

I/O

Trait	Methods	Format
`CSV`	`write_csv(path, delimiter)`, `read_csv(path, delimiter)`	CSV
`JsonIO`	`write_json(path)`, `read_json(path)`, `to_json_string()`, `from_json_string(s)`	JSON

JSON format:

{
  "headers": ["x", "y"],
  "data": {
    "x": [1, 2, 3],
    "y": [4.0, 5.0, 6.0]
  }
}

Supported Column Types

Vec<bool>, Vec<u32>, Vec<u64>, Vec<usize>, Vec<i32>, Vec<i64>, Vec<isize>, Vec<f32>, Vec<f64>, Vec<String>, Vec<&str>

Any type implementing the Column trait can be used as a column.

License

Licensed under either of

Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
MIT License (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.

puruda 0.2.0