db-migrator

A fast, type-safe MSSQL to MySQL database migration tool written in Rust.

Features • Quick Start • Installation • Configuration • Usage • Type Mappings

---

$ db-migrator --format --constraints --verbose

Migrating table: user_accounts
Table user_accounts migrated, rows: 12840, took: 1.23s
Migrating table: orders
Table orders migrated, rows: 58201, took: 3.47s
Migration finished, total rows: 71041, total time took: 4.82s

Features

• Schema + data migration — transfers table structures, column types, constraints, and all rows
• 29 MSSQL types mapped — built-in defaults for every type, zero config needed
• Concurrent — tables migrate in parallel with configurable parallelism
• Batch inserts — rows streamed and batched to respect MySQL max_allowed_packet
• Constraints — primary keys and foreign keys carried over automatically
• Snake case — optionally convert PascalCase names to snake_case
• Customizable — override any type mapping with a simple config file

Quick Start

# 1. Configure your databases
cp config.example.toml config.toml    # edit connection details

# 2. Run
cargo run --release

That's it. All type mappings are built in — no extra config needed unless you want to customize them.

Installation

Pre-compiled Binaries

Download the latest binary for your platform from Releases.

Build from Source

git clone https://github.com/bitalizer/db-migrator.git
cd db-migrator
cargo build --release

Requires Rust 1.85+ (edition 2024).

Configuration

Database Connections

Create config.toml from the example:

[mssql_database]
host = "localhost"   # optional, defaults shown
port = 1433
username = "db_user"
password = "db_pass"
database = "source_db"

[mysql_database]
host = "localhost"   # optional, defaults shown
port = 3306
username = "db_user"
password = "db_pass"
database = "target_db"

[settings]
max_packet_bytes = 1048576
whitelisted_tables = ["Users", "Orders", "Products"]

Character sets: text is decoded from SQL Server to UTF-8 during extraction, and created tables inherit the target database's default charset. Create the target database with utf8mb4 (CREATE DATABASE output CHARACTER SET utf8mb4) so all characters survive. Collation semantics are not translated: data from a case-sensitive SQL Server column can collide on MySQL's default case-insensitive unique indexes.

Usage

db-migrator [FLAGS] [OPTIONS]

Flag	Description
-c, --constraints	Create primary key and foreign key constraints
-d, --drop	Drop target tables before migration
-f, --format	Convert table and column names to snake_case
-v, --verbose	Verbose logging
-q, --quiet	Suppress output

Option	Description
--source <URL>	Source database URL, mssql://user:pass@host:1433/database
--target <URL>	Target database URL, mysql://user:pass@host:3306/database
--tables <T1,T2>	Comma-separated tables to migrate (CLI mode)
-p, --parallelism <N>	Max concurrent table migrations (default: CPU cores)
--max-packet-bytes <N>	Max INSERT batch size in bytes (default: 1048576; overrides config.toml)

CLI mode

Passing any of --source/--target/--tables skips config.toml entirely, so a migration is a single copy-pasteable command. All three are required in this mode; the port defaults to the engine port (1433/3306) when omitted from a URL.

db-migrator \
  --source "mssql://sa:pass@10.0.0.1:1433/input" \
  --target "mysql://root:pass@localhost:3306/output" \
  --tables Users,Orders,Products \
  --format --constraints

Credentials passed on the command line are visible in shell history and process lists. Percent-encode special characters in passwords (p@ss becomes p%40ss). mappings.toml type overrides apply in both modes.

Examples

# Basic migration
db-migrator

# Full migration with snake_case and constraints
db-migrator --format --constraints

# Verbose with custom parallelism
db-migrator -v -p 4

# Drop and recreate all tables
db-migrator --drop --constraints --format

Type Mappings

All 29 MSSQL types are mapped by default. No configuration required.

View full type mapping table

MSSQL	MySQL	Notes
bit	tinyint(1)	Boolean equivalent
tinyint	tinyint unsigned
smallint	smallint
int	int
bigint	bigint
decimal / numeric	decimal	Precision and scale carried over
money	decimal(19, 4)
smallmoney	decimal(10, 2)
float	float
real	real
char / nchar	char	Length carried from source
varchar	varchar	Length carried, default 255
nvarchar	longtext	Unicode safe
text / ntext	longtext
binary	binary	Length carried from source
varbinary / image	longblob
date	date
datetime / datetime2 / smalldatetime	datetime
datetimeoffset	datetime	Offset stripped
time	time
uniqueidentifier	char(36)	UUID as string
timestamp	bigint unsigned	Row version counter, not a time value
xml	longtext

Custom Overrides (optional)

To override any default mapping, create a mappings.toml file in the project root:

[mappings]
nvarchar = "varchar(500)"
money = "decimal(10, 2)"
float = "float(53)"
xml = "longtext"

Supports three formats: "type", "type(length)", and "type(precision, scale)", each optionally followed by the modifiers unsigned and/or zerofill (numeric types only), e.g. "int unsigned" or "decimal(19, 4) unsigned".

Column-scoped overrides

Type-wide overrides apply to every column of that source type. When only specific columns need different output (e.g. a legacy schema expecting int unsigned IDs), scope the override to "Table.Column" source names:

[mappings.columns]
"Orders.ID" = "int unsigned"
"Orders.Notes" = "text"

Column overrides take precedence over type-wide overrides, which take precedence over the built-in defaults. Names are matched case-insensitively against the source (SQL Server) table and column names.

Requirements

• Source: MSSQL Server
• Target: MySQL 5.7+ / MariaDB 10.3+
• Build: Rust 1.85+

Dependencies

Crate	Purpose
tokio	Async runtime
tiberius	MSSQL driver
sqlx	MySQL driver
bb8	Connection pooling
clap	CLI parsing
anyhow	Error handling
chrono	Date/time
toml	Config parsing

License

See LICENSE for details.