SawitDB

SawitDB is a unique database solution stored in .sawit binary files.

The system features a custom Hybrid Paged Architecture similar to SQLite but supercharged with Object Caching, using fixed-size 4KB pages to ensure efficient memory usage and near-instant access. What differentiates SawitDB is its unique Agricultural Query Language (AQL), which replaces standard SQL keywords with Indonesian farming terminology.

Now available on NPM! Connect via TCP using sawitdb:// protocol.

🚨 Emergency: Aceh Flood Relief Please support our brothers and sisters in Aceh.

Organized by Human Initiative Aceh

Features

• Hybrid Paged Architecture: Data is stored in 4096-byte binary pages, but hot data is cached as native Objects for zero-copy reads.
• Single File Storage: All data, schema, and indexes are stored in a single .sawit file.
• High Stability: Uses 4KB atomic pages. More stable than a coalition government.
• Data Integrity (Anti-Korupsi): Implements strict fsync protocols. Data cannot be "corrupted" or "disappear" mysteriously like social aid funds (Bansos). No "Sunat Massal" here.
• Crash Recovery: Uses Write-Ahead Logging (WAL). Guarantees data always returns after a crash. Unlike a fugitive (Buronan) who is "hard to find".
• Zero Bureaucracy (Zero Deps): Built entirely with standard Node.js. No unnecessary "Vendor Pengadaan" or "Mark-up Anggaran".
• Transparansi: Query language is clear. No "Pasal Karet" (Ambiguous Laws) or "Rapat Tertutup" in 5-star hotels.
• Speed: Faster than printing an e-KTP at the Kelurahan.
• Network Support (NEW): Client-Server architecture with Multi-database support and Authentication.
• True Multi-Threading (NEW): Worker Pool architecture separates IO (Main Thread) from CPU (Worker Threads).
• Advanced SQL (NEW): Support for JOIN (Left/Right/Full/Cross), JAVING, DISTINCT, and more.
• NPM Support (NEW): Install via npm install @wowoengine/sawitdb.

Filosofi

Filosofi (ID)

SawitDB dibangun dengan semangat "Kemandirian Data". Kami percaya database yang handal tidak butuh Infrastruktur Langit yang harganya triliunan tapi sering down. Berbeda dengan proyek negara yang mahal di budget tapi murah di kualitas, SawitDB menggunakan arsitektur Single File (.sawit) yang hemat biaya. Backup cukup copy-paste, tidak perlu sewa vendor konsultan asing. Fitur fsync kami menjamin data tertulis di disk, karena bagi kami, integritas data adalah harga mati, bukan sekadar bahan konferensi pers untuk minta maaf.

Philosophy (EN)

SawitDB is built with the spirit of "Data Sovereignty". We believe a reliable database doesn't need "Sky Infrastructure" that costs trillions yet goes down often. Unlike state projects that are expensive in budget but cheap in quality, SawitDB uses a cost-effective Single File (.sawit) architecture. Backup is just copy-paste, no need to hire expensive foreign consultants. Our fsync feature guarantees data is written to disk, because for us, data integrity is non-negotiable, not just material for a press conference to apologize.

File List

• src/WowoEngine.js: Core Database Engine (Class: SawitDB).
• src/SawitServer.js: Server Class (TCP/Auth/ThreadPool).
• src/SawitClient.js: Client Class.
• src/modules/ThreadPool.js: Worker Pool Manager.
• src/SawitWorker.js: Worker Entry Point.
• bin/sawit-server.js: Server executable.
• cli/local.js: Interactive CLI tool (Local).
• cli/remote.js: Interactive CLI tool (Network).
• cli/test.js: Unit Test Suite.
• cli/benchmark.js: Performance Benchmark Tool.
• CHANGELOG.md: Version history.

Installation

Install via NPM:

npm install @wowoengine/sawitdb

Quick Start

1. Start the Server

node bin/sawit-server.js
# Or with Cluster Mode enabled in .env

The server will start on 0.0.0.0:7878 by default.

2. Connect with Client

You can use the built-in CLI tool:

node cli/remote.js

Or use the SawitClient class in your Node.js application.

---

Dual Syntax Support

SawitDB introduces the Generic Syntax alongside the classic Agricultural Query Language (AQL), making it easier for developers familiar with standard SQL to adopt.

Operation	Agricultural Query Language (AQL)	Generic SQL (Standard)
Create DB	BUKA WILAYAH sales_db	CREATE DATABASE sales_db
Use DB	MASUK WILAYAH sales_db	USE sales_db
Show DBs	LIHAT WILAYAH	SHOW DATABASES
Drop DB	BAKAR WILAYAH sales_db	DROP DATABASE sales_db
Create Table	LAHAN products	CREATE TABLE products
Insert	TANAM KE products (...) BIBIT (...)	INSERT INTO products (...) VALUES (...)
Select	PANEN * DARI products DIMANA ...	SELECT * FROM products WHERE ...
Update	PUPUK products DENGAN ...	UPDATE products SET ...
Delete	GUSUR DARI products DIMANA ...	DELETE FROM products WHERE ...
Indexing	INDEKS products PADA price	CREATE INDEX ON products (price)
Aggregation	HITUNG SUM(stock) DARI products	Same Syntax

---

Query Syntax (Detailed)

1. Management Commands

Create Table

-- Tani
LAHAN users
-- Generic
CREATE TABLE users

Show Tables

-- Tani
LIHAT LAHAN
-- Generic
SHOW TABLES

Drop Table

-- Tani
BAKAR LAHAN users
-- Generic
DROP TABLE users

2. Data Manipulation

Insert Data

-- Tani
TANAM KE users (name, role) BIBIT ('Alice', 'Admin')
-- Generic
INSERT INTO users (name, role) VALUES ('Alice', 'Admin')

Select Data

-- Tani
PANEN name, role DARI users DIMANA role = 'Admin' ORDER BY name ASC LIMIT 10
-- Generic
SELECT name, role FROM users WHERE role = 'Admin' ORDER BY name ASC LIMIT 10

Operators: =, !=, >, <, >=, <= Advanced: IN ('a','b'), LIKE 'pat%', BETWEEN 10 AND 20, IS NULL, IS NOT NULL

Pagination & Sorting

SELECT * FROM users ORDER BY age DESC LIMIT 5 OFFSET 10
SELECT * FROM users WHERE age BETWEEN 18 AND 30 AND status IS NOT NULL

Update Data

-- Tani
PUPUK users DENGAN role='SuperAdmin' DIMANA name='Alice'
-- Generic
UPDATE users SET role='SuperAdmin' WHERE name='Alice'

Delete Data

-- Tani
GUSUR DARI users DIMANA name='Bob'
-- Generic
DELETE FROM users WHERE name='Bob'

3. Advanced Features

Indexing

INDEKS [table] PADA [field]
-- or
CREATE INDEX ON [table] ([field])

Aggregation & Grouping

HITUNG COUNT(*) DARI [table]
HITUNG AVG(price) DARI [products] KELOMPOK [category]
-- With HAVING clause
HITUNG COUNT(*) DARI sales GROUP BY region HAVING count > 5

DISTINCT

SELECT DISTINCT category FROM products
-- Returns only unique values

JOIN Types

-- INNER JOIN (default)
SELECT * FROM orders JOIN customers ON orders.customer_id = customers.id

-- LEFT OUTER JOIN
SELECT * FROM employees LEFT JOIN departments ON employees.dept_id = departments.id

-- RIGHT OUTER JOIN
SELECT * FROM employees RIGHT JOIN departments ON employees.dept_id = departments.id

-- CROSS JOIN (Cartesian product)
SELECT * FROM colors CROSS JOIN sizes

EXPLAIN Query Plan

EXPLAIN SELECT * FROM users WHERE id = 5
-- Returns execution plan: scan type, index usage, join methods

Architecture Details

• Worker Pool (Multi-threaded):
  • Main Thread: Handles strictly I/O (Networking, Protocol Parsing).
  • Worker Threads: Execute queries in parallel using Least-Busy Load Balancing.
  • Fault Tolerance: Automatic worker healing and stuck-query rejection.
• Storage Engine:
  • Hybrid Paging: 4KB binary pages with In-Memory Object Cache for hot data.
  • WAL (Write-Ahead Log): Ensures ACID properties and crash recovery.
  • B-Tree Indexing: O(log n) lookups for high performance.
• Modular Codebase: Core logic separated into src/modules/ (Pager.js, ThreadPool.js, BTreeIndex.js).

Benchmark Performance

Test Environment: Single Thread, Windows Node.js (Local NVMe)

Operation	Ops/Sec	Latency (avg)
INSERT	~22,000	0.045 ms
SELECT (PK Index)	~247,288	0.004 ms
SELECT (Scan)	~13,200 (10k rows)	0.075 ms
UPDATE (Indexed)	~11,000	0.090 ms
DELETE (Indexed)	~19,000	0.052 ms

Note: Hasil diatas adalah benchmark Internal (Engine-only). Untuk Network Benchmark (Cluster Mode), TPS berkisar ~20.000 (overhead jaringan).

Full Feature Comparison

Feature	Tani Edition (AQL)	Generic SQL (Standard)	Notes
Create DB	BUKA WILAYAH [db]	CREATE DATABASE [db]	Creates .sawit in data/
Use DB	MASUK WILAYAH [db]	USE [db]	Switch context
Show DBs	LIHAT WILAYAH	SHOW DATABASES	Lists available DBs
Drop DB	BAKAR WILAYAH [db]	DROP DATABASE [db]	Irreversible!
Create Table	LAHAN [table]	CREATE TABLE [table]	Schema-less creation
Show Tables	LIHAT LAHAN	SHOW TABLES	Lists tables in DB
Drop Table	BAKAR LAHAN [table]	DROP TABLE [table]	Deletes table & data
Insert	TANAM KE [table] ... BIBIT (...)	INSERT INTO [table] (...) VALUES (...)	Auto-ID if omitted
Select	PANEN ... DARI [table] DIMANA ...	SELECT ... FROM [table] WHERE ...	Supports Projection
Update	PUPUK [table] DENGAN ... DIMANA ...	UPDATE [table] SET ... WHERE ...	Atomic update
Delete	GUSUR DARI [table] DIMANA ...	DELETE FROM [table] WHERE ...	Row-level deletion
Index	INDEKS [table] PADA [field]	CREATE INDEX ON [table] (field)	B-Tree Indexing
Count	HITUNG COUNT(*) DARI [table]	SELECT COUNT(*) FROM [table] (via HITUNG)	Aggregation
Sum	HITUNG SUM(col) DARI [table]	SELECT SUM(col) FROM [table] (via HITUNG)	Aggregation
Average	HITUNG AVG(col) DARI [table]	SELECT AVG(col) FROM [table] (via HITUNG)	Aggregation
Min/Max	HITUNG MIN(col) DARI [table]	SELECT MIN(col) FROM [table] (via HITUNG)	Aggregation
DISTINCT	PANEN UNIK col DARI [table]	SELECT DISTINCT col FROM [table]	Unique rows
LEFT JOIN	GABUNG KIRI	LEFT JOIN ... ON ...	Outer join
RIGHT JOIN	GABUNG KANAN	RIGHT JOIN ... ON ...	Outer join
CROSS JOIN	GABUNG SILANG	CROSS JOIN	Cartesian product
HAVING	PUNYA count > 5	HAVING count > 5	Filter groups
EXPLAIN	JELASKAN SELECT ...	EXPLAIN SELECT ...	Query plan

Supported Operators Table

Operator	Syntax Example	Description
Comparison	=, !=, >, <, >=, <=	Standard value comparison
Logical	AND, OR	Combine multiple conditions
In List	IN ('coffee', 'tea')	Matches any value in the list
Not In	NOT IN ('water')	Matches values NOT in list
Pattern	LIKE 'Jwa%'	Standard SQL wildcard matching
Range	BETWEEN 1000 AND 5000	Inclusive range check
Null	IS NULL	Check if field is empty/null
Not Null	IS NOT NULL	Check if field has value
Distinct	SELECT DISTINCT col	Remove duplicate rows
Limit	LIMIT 10	Restrict number of rows
Offset	OFFSET 5	Skip first N rows (Pagination)
Order	ORDER BY price DESC	Sort by field (ASC/DESC)

License

MIT License