CSV to SQL Converter

How to use it

1. Paste CSV or JSON, or drop a file — a badge names the detected format; one click corrects it.
2. Pick dialect and table name, then check the column grid — override any inferred type per column.
3. Set batch size, empty-cell handling and identifier case; DROP TABLE and TRUNCATE need their own confirmation.
4. Copy the preview or download the whole script as .sql — the preview stops at 100 statements, the download never does.

Seeding a database from an export usually means hand-writing INSERT statements around every apostrophe. Paste CSV or JSON rows and get CREATE TABLE plus batched INSERTs for PostgreSQL, MySQL, SQLite or SQL Server — column types inferred from every row, identifiers quoted where the dialect demands it, string values escaped the way that dialect actually reads them. Nothing you paste leaves this device.

Examples

id,name,city,rank
1,O'Brien,Cork,3
2,Ana,Lisbon,1
3,Löwe,Berlin,2
4,Sam,Leeds,4

CREATE TABLE customers (
  id INTEGER,
  name TEXT,
  city TEXT,
  rank INTEGER
);
INSERT INTO customers (id, name, city, rank) VALUES
  (1, 'O''Brien', 'Cork', 3),
  (2, 'Ana', 'Lisbon', 1);
INSERT INTO customers (id, name, city, rank) VALUES
  (3, 'Löwe', 'Berlin', 2),
  (4, 'Sam', 'Leeds', 4);

name,city,rank,id
O'Brien,Cork,3,1
Ana,Lisbon,1,2
Löwe,Berlin,2,3
Sam,Leeds,4,4

CREATE TABLE customers (
  name TEXT,
  city TEXT,
  `rank` INT,
  id INT
);
INSERT INTO customers (name, city, `rank`, id) VALUES
  ('O\'Brien', 'Cork', 3, 1),
  ('Ana', 'Lisbon', 1, 2),
  ('Löwe', 'Berlin', 2, 3),
  ('Sam', 'Leeds', 4, 4);

[{"id":1,"name":"Ada","tags":["maps","sql"]},{"id":2,"city":"Oslo"}]

CREATE TABLE people (
  id INTEGER,
  name TEXT,
  tags TEXT,
  city TEXT
);
INSERT INTO people (id, name, tags, city) VALUES
  (1, 'Ada', '["maps","sql"]', NULL),
  (2, NULL, NULL, 'Oslo');

sku,price,in_stock

CREATE TABLE products (
  sku TEXT,
  price TEXT,
  in_stock TEXT
);

Every example on this page is executed against the same engine that powers the tool, on every build.

Common errors

Input is larger than the 100 MB limit for this tool.

Refusal, not truncation — a script silently missing rows would look finished. At that size, the sqlite3 line below imports the file directly.

Invalid JSON: Expected property name or '}' at line 1, column 14 — fix the syntax there, or set the format to CSV if this is delimited data.

{id: 2} is a JavaScript object literal, not JSON — keys need double quotes. The position marks the first offending key.

JSON array element 1 is a number — every row must be an object like {"column": value}, so the rows share named columns.

Bare values carry no column names, and INSERT tuples need them. Wrap each value in an object, or paste CSV with a header row.

DROP TABLE is enabled but not confirmed — tick the destructive confirmation to include it, or turn the toggle off.

Destructive statements are never included silently. The confirmation checkbox beside those toggles is what this message asks for.

Quoted field opened at line 2, column 3 never closes — this usually breaks every following row; close the quote, or write a literal quote inside a quoted field as "".

One unclosed quote makes every later row read as the same cell — add the closing quote where the position points, or double literal quotes inside the field.

Keyboard shortcuts

⌘/Ctrl + Enter

Convert again now — above 2 MB this is the explicit Run.

⌘/Ctrl + S

Download the full script as .sql.

⌘/Ctrl + C

Copy the script when nothing is selected.

⌘/Ctrl + Shift + C

Copy even while a selection exists.

t

Cycle the theme.

? or ⌘/Ctrl + /

Show the shortcut help overlay.

Esc

Close the topmost overlay.