VSL CSV Module

The vsl.inout.csv module provides utilities for reading and writing CSV files commonly used in scientific computing and machine learning workflows.

Features

• CSV Reading: Parse CSV files into numeric arrays or matrices
• CSV Writing: Export data to CSV format
• Flexible Configuration: Handle headers, delimiters, comments, and column selection
• Matrix Integration: Direct conversion to vsl.la.Matrix types

Quick Start

Reading CSV Files

import vsl.inout.csv

// Basic CSV reading
csv_data := csv.read_csv('data.csv', csv.CsvReadConfig{})!
println('Rows: ${csv_data.n_rows}, Cols: ${csv_data.n_cols}')

// Access the data
for row in csv_data.data {
    println(row)
}

Reading with Header

import vsl.inout.csv

csv_data := csv.read_csv('data.csv', csv.CsvReadConfig{
    skip_header: true
})!

println('Columns: ${csv_data.header}')

// Get column by name
ages := csv_data.get_column_by_name('age')!

Reading Specific Columns

import vsl.inout.csv

// Only read columns 0, 2, and 4
csv_data := csv.read_csv('data.csv', csv.CsvReadConfig{
    use_cols: [0, 2, 4]
})!

Converting to Matrix

import vsl.inout.csv

// Read directly as a Matrix
matrix := csv.read_csv_to_matrix('data.csv', csv.CsvReadConfig{
    skip_header: true
})!

println('Matrix dimensions: ${matrix.m} x ${matrix.n}')

Writing CSV Files

import vsl.inout.csv

data := [[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]]

csv.write_csv('output.csv', data, csv.CsvWriteConfig{
    header: ['x', 'y', 'z']
})!

Configuration Options

CsvReadConfig

CsvWriteConfig

CsvData Methods

• to_matrix() - Convert to la.Matrix[f64]
• get_column(idx) - Get column by index
• get_column_by_name(name) - Get column by header name

Examples

Handling Different Delimiters

import vsl.inout.csv

// Tab-separated values
csv_data := csv.read_csv('data.tsv', csv.CsvReadConfig{
    delimiter: `\t`
})!

// Semicolon-separated
csv_data2 := csv.read_csv('data.csv', csv.CsvReadConfig{
    delimiter: `;`
})!

Parsing CSV String

import vsl.inout.csv

content := 'x,y
1,2
3,4
5,6'

csv_data := csv.parse_csv(content, csv.CsvReadConfig{
    skip_header: true
})!

Integration with ML Module

import vsl.inout.csv
import vsl.ml

// Read data for machine learning
csv_data := csv.read_csv('dataset.csv', csv.CsvReadConfig{
    skip_header: true
})!

// Create ML Data object
matrix := csv_data.to_matrix()!
mut ml_data := ml.Data.from_raw_xy(csv_data.data)!

// Split into train/test
train_data, test_data := ml_data.split(0.8)!

See Also

• HDF5 Module
• Data Preparation
• Linear Algebra
• Machine Learning

fn parse_csv(content string, config CsvReadConfig) !CsvData

parse_csv parses CSV content from a string

fn read_csv(path string, config CsvReadConfig) !CsvData

read_csv reads a CSV file and returns numeric data as a 2D array Non-numeric values are converted to NaN

fn read_csv_to_matrix(path string, config CsvReadConfig) !&la.Matrix[f64]

read_csv_to_matrix reads a CSV file and returns the data as a Matrix

fn write_csv(path string, data [][]f64, config CsvWriteConfig) !

write_csv writes a 2D array to a CSV file

fn write_matrix_csv(path string, matrix &la.Matrix[f64], config CsvWriteConfig) !

write_matrix_csv writes a Matrix to a CSV file

struct CsvData {
pub:
	data        [][]f64    // the numeric data
	header      []string   // header row (if present)
	n_rows      int        // number of data rows
	n_cols      int        // number of columns
	raw_strings [][]string // original string data (for non-numeric columns)
}

CsvData holds the result of reading a CSV file

fn (c &CsvData) to_matrix() !&la.Matrix[f64]

to_matrix converts CsvData to a Matrix

fn (c &CsvData) get_column(idx int) ![]f64

get_column returns a specific column as an array

fn (c &CsvData) get_column_by_name(name string) ![]f64

get_column_by_name returns a column by its header name

struct CsvReadConfig {
pub:
	delimiter   u8 = `,` // field delimiter
	skip_header bool  // skip the first row
	skip_rows   int   // number of rows to skip at the beginning
	use_cols    []int // indices of columns to use (empty = all)
	max_rows    int = -1  // maximum number of rows to read (-1 = all)
	comment     u8  = `#` // comment character (lines starting with this are skipped)
}

CsvReadConfig holds configuration for reading CSV files

struct CsvWriteConfig {
pub:
	delimiter    u8 = `,` // field delimiter
	header       []string // optional header row
	float_format string = '%.6g' // format for floating point numbers
}

CsvWriteConfig holds configuration for writing CSV files