slowlog

A Rust library for parsing and analysing MySQL slow query logs. This library anonymises SQL queries by replacing literal values with placeholders, making it easy to identify and group similar queries for performance analysis.

Features

• Parse MySQL slow query log files
• Anonymise queries by replacing literals with placeholders
• Extract detailed query statistics (query time, lock time, rows examined, etc.)
• Generate fingerprints for normalised queries

Installation

cargo add slowlog

Usage

Parsing Slow Log Files

use slowlog::process_slow_log_file;

fn main() {
    process_slow_log_file("path/to/slow.log", |query| {
        println!("Original: {}", query.query);
        println!("Formatted: {}", query.formatted);
        println!("Fingerprint: {}", query.fingerprint);
        println!("Query time: {:.2}s", query.stats.query_time);
        println!("Rows examined: {}", query.stats.rows_examined);
    }).unwrap();
}

Processing from a Reader

use slowlog::process_slow_log_reader;
use std::fs::File;
use std::io::BufReader;

fn main() {
    let file = File::open("slow.log").unwrap();
    let reader = BufReader::new(file);

    process_slow_log_reader(reader, |query| {
        println!("{:?}", query);
    }).unwrap();
}

Query Anonymisation Examples

The library replaces all literal values with ? placeholders:

Query Statistics

The library extracts the following statistics from slow log entries:

• user: Database user who executed the query
• host: Host from which the query was executed
• time: Timestamp when the query was executed
• query_time: Total query execution time in seconds
• lock_time: Time spent waiting for locks in seconds
• rows_sent: Number of rows returned by the query
• rows_examined: Number of rows examined during query execution

API Documentation

Functions

process_slow_log_file<Q>(path: &str, query_callback: Q) -> io::Result<()>

Processes a slow log file and calls the callback for each query found.

Parameters:

• path: Path to the slow log file
• query_callback: Function called for each parsed query

Example:

use slowlog::process_slow_log_file;

process_slow_log_file("slow.log", |query| {
    println!("{:?}", query);
}).unwrap();

process_slow_log_reader<R, Q>(reader: R, query_callback: Q) -> io::Result<()>

Processes slow log data from any BufRead source.

Parameters:

• reader: Any type implementing BufRead
• query_callback: Function called for each parsed query

Example:

use slowlog::process_slow_log_reader;
use std::io::BufReader;

let reader = BufReader::new(file);
process_slow_log_reader(reader, |query| {
    println!("{:?}", query);
}).unwrap();

Types

Query

pub struct Query {
    pub query: String,        // Original query
    pub formatted: String,    // Anonymised query with placeholders
    pub fingerprint: String,  // SHA1 hash of formatted query
    pub stats: QueryStats,    // Query execution statistics
}

QueryStats

pub struct QueryStats {
    pub user: String,           // Database user
    pub host: String,           // Client host
    pub time: DateTime<Utc>,    // Execution timestamp
    pub rows_examined: u64,     // Rows scanned
    pub rows_sent: u64,         // Rows returned
    pub query_time: f64,        // Execution time (seconds)
    pub lock_time: f64,         // Lock wait time (seconds)
}

QueryError

pub enum QueryError {
    ParseError(String),  // SQL parsing failed
    InvalidQuery,        // No valid SQL statement found
}

Supported SQL Features

• Basic SELECT, INSERT, UPDATE, DELETE statements
• WHERE clauses with operators (=, !=, <>, >, <, >=, <=)
• IN and NOT IN lists
• BETWEEN conditions
• LIKE and NOT LIKE patterns
• Subqueries
• CASE expressions
• SQL functions (MAX, MIN, CEIL, FLOOR, etc.)
• GROUPING operations
• Binary and unary operations

License

MIT

Repository

https://github.com/tpyo/slowlog

Author

Don Schonknecht don@tpyo.net