SQL to JSON Converter

Convert SQL to JSON

This tool extracts data from SQL INSERT statements and converts it to JSON format. It parses column names and values, handling strings, numbers, booleans, and NULL values.

Supported SQL Syntax

The converter handles:

• INSERT INTO with explicit column names
• Multiple value sets in single INSERT
• Multiple INSERT statements
• CREATE TABLE for column name extraction
• Standard SQL value types (strings, numbers, booleans, NULL)

Example Conversion

Input SQL:

INSERT INTO users (id, name, active) VALUES
  (1, 'Alice', TRUE),
  (2, 'Bob', FALSE);

Output JSON:

[
  { "id": 1, "name": "Alice", "active": true },
  { "id": 2, "name": "Bob", "active": false }
]

Type Conversion

Common Use Cases

• Database exports — Convert SQL dumps to JSON for analysis
• API migration — Transform database data for REST APIs
• Testing — Generate JSON fixtures from SQL test data
• Documentation — Convert sample data to readable JSON

Handling Edge Cases

Escaped quotes

Single quotes within strings should be escaped as '' (standard SQL escaping):

INSERT INTO messages (text) VALUES ('It''s working');
-- Becomes: { "text": "It's working" }

Multiple tables

If your SQL contains inserts to multiple tables, all rows are combined. For best results, convert one table at a time.

Programmatic Conversion

Python

import re
import json

def sql_insert_to_json(sql):
    # Simple regex for INSERT VALUES
    pattern = r"INSERT INTO w+ (([^)]+)) VALUES (.+)"
    match = re.search(pattern, sql, re.IGNORECASE)
    
    columns = [c.strip() for c in match.group(1).split(',')]
    # Parse values and build objects...
    return json.dumps(result, indent=2)

JavaScript

// Using a SQL parser library
import { Parser } from 'node-sql-parser';

const parser = new Parser();
const ast = parser.astify(sql);
// Extract data from AST...

Related Tools

• JSON to SQL — Convert JSON back to SQL
• JSON Validator — Validate the output
• JSON to CSV — Export as spreadsheet
• CSV to JSON — Alternative data import

Frequently Asked Questions

Can I convert SELECT query results?

This tool parses INSERT statements, not query results. For SELECT output, export as CSV first, then use our CSV to JSON converter.

What SQL dialects are supported?

The parser handles standard SQL syntax compatible with MySQL, PostgreSQL, SQLite, and SQL Server. Dialect-specific features may not be supported.

How do I handle large SQL files?

The tool runs in your browser, so very large files may be slow. For massive datasets, consider using command-line tools or programming libraries.

Common Mistakes & Pro Tips

• Column names come from the INSERT's column list — JSON keys are taken from the (col1, col2, ...) list in INSERT INTO table (col1, col2) VALUES (...). If that list is omitted, the converter can only fall back to positional keys like "column_1" because the column names live in the table definition, not the statement.
• SQL NULL is not the string 'NULL' — An unquoted NULL token must become JSON null, while the quoted literal 'NULL' is the text "NULL". Mixing these up is a common bug, so confirm that bare NULL in your VALUES list maps to null and only quoted values become strings.
• Escaped quotes inside string values — In standard SQL a single quote inside a string is escaped by doubling it ('O''Brien'), while some dumps (notably MySQL) use a backslash (\'). The parser must unescape these back to one quote — O'Brien — rather than leaving the doubled or backslashed form in the JSON.
• Numbers, booleans, and dates are dialect-dependent — Unquoted numerics like 42 or 3.14 become JSON numbers, but booleans vary: some databases use TRUE/FALSE, others 0/1, and dates are usually quoted strings like '2024-01-01'. Don't assume a column is a real boolean or date type just from the literal.
• Multi-row INSERTs and trailing semicolons — A single INSERT can list many VALUES (...) tuples separated by commas, each becoming one object, and statements end with a semicolon. Comments (-- or /* */) and other DDL/DML around the inserts are ignored, so only the INSERT data is extracted.