JSON to SQL Converter

Paste a JSON array of objects into the input panel. The converter analyzes the keys and value types across all objects to infer column names and SQL data types. It generates a CREATE TABLE statement with appropriate column types (VARCHAR for strings, INTEGER or NUMERIC for numbers, BOOLEAN, TIMESTAMP for date strings) and INSERT INTO statements for each row. You can select your target database dialect (PostgreSQL, MySQL, or SQLite) since they differ in type names, quoting rules, and syntax. String values are automatically escaped to prevent SQL injection in the generated output. Null JSON values produce SQL NULL literals in the corresponding INSERT columns.

PostgreSQL uses rich types like TEXT (unlimited strings), INTEGER, BIGINT, NUMERIC (precise decimals), BOOLEAN, TIMESTAMP WITH TIME ZONE, JSONB (indexed JSON), and UUID. MySQL uses VARCHAR(255) for strings (requiring length), INT, BIGINT, DECIMAL, TINYINT(1) for booleans, DATETIME, JSON, and CHAR(36) for UUIDs. SQLite has only five storage classes: TEXT, INTEGER, REAL, BLOB, and NULL with dynamic typing that accepts any value in any column. The converter maps JSON types to the most appropriate native type for each dialect. Numbers without decimals become INTEGER, numbers with decimals become NUMERIC or DECIMAL, and ISO date strings become TIMESTAMP or DATETIME.

Nested JSON objects and arrays cannot be directly mapped to flat SQL columns in traditional relational fashion. The converter offers two approaches. The first serializes nested values as JSON strings stored in a JSON or TEXT column, preserving the full structure. PostgreSQL's JSONB and MySQL's JSON types support querying into these values. The second approach flattens one level of nesting by creating separate columns with dot-notation names like address_city and address_zip. Arrays of primitive values are serialized as JSON strings. For deeply nested structures, consider normalizing the data into multiple related tables manually, as automated flattening beyond one level produces unwieldy schemas.