How does column type inference work?

The converter examines every non-empty value in a column. If all are whole numbers it picks integer, fractional numbers give decimal, true/false/yes/no values give boolean, recognizable dates (YYYY-MM-DD, MM/DD/YYYY, or DD-MM-YYYY) give date, and otherwise it uses VARCHAR sized to the longest value (50 to 255 characters) or TEXT when values exceed 255. You can override any inferred type per column from the preview table.

How does the dialect setting change the output?

Only the SQL syntax changes; the data is the same. PostgreSQL uses double-quoted identifiers and TRUE/FALSE booleans. MySQL uses backtick identifiers and TINYINT(1) with 1/0 booleans. SQL Server uses bracketed identifiers, NVARCHAR, and BIT. SQLite uses INTEGER and REAL with loose typing. The right column types and boolean literals are emitted for whichever dialect you pick, so the script loads without edits.

Does it handle quoted CSV fields, commas, and newlines inside values?

Yes. The parser is RFC-4180 friendly: double-quoted fields keep commas and even newlines as part of the value, and a doubled double-quote ("") inside a quoted field is treated as a single escaped quote. The delimiter is auto-detected from the header line, or you can force comma, semicolon, tab, or pipe.

How are empty cells and special characters handled?

Empty cells become NULL in the generated INSERTs, and columns that contain any empty cell are left nullable in the CREATE TABLE (no NOT NULL). Text values have their single quotes escaped by doubling them, so a value like O'Brien is written safely as 'O''Brien'.

Can I generate only the schema or only the data?

Yes. The Include CREATE TABLE and Include INSERT statements toggles are independent, so you can produce just the schema, just the inserts, or both. For the inserts you can choose one multi-row INSERT per batch (with a configurable batch size) or one INSERT statement per row.

What if my CSV has no header row?

Turn off "First row is a header" and the converter treats every row as data, naming the columns column_1, column_2, and so on. You can still override each column type and rename the target table; only the column identifiers are auto-generated.

CSV to SQL Converter

Convert CSV into ready-to-run CREATE TABLE and INSERT statements. The parser handles quoted fields with embedded commas and newlines, auto-detects the delimiter, and infers a SQL type for each column (with a per-column override). Pick PostgreSQL, MySQL, SQLite, or SQL Server and the identifiers, column types, and boolean literals adapt to your dialect. Live output updates as you type - copy, download, or share a link.

About this ToolHow it works, benefits & use cases

Turn a block of CSV into a ready-to-run table and the INSERT statements to populate it, with output that adapts to your database. Paste your CSV and the converter auto-detects the delimiter (or you can pick comma, semicolon, tab, or pipe), reads the header row, sanitizes each name into a safe column identifier, and infers a SQL type per column: integer for whole numbers, decimal for fractional numbers, boolean for true/false/yes/no values, date for recognizable date formats, and VARCHAR (sized to the longest value) or TEXT for longer strings. Every inferred type can be overridden per column from the preview table, and empty cells become NULL. Choose PostgreSQL, MySQL, SQLite, or SQL Server and the identifier quoting, column types, and boolean literals all change to match - double quotes and TRUE/FALSE for Postgres, backticks and TINYINT(1) for MySQL, brackets and BIT for SQL Server. Toggle the CREATE TABLE header and the INSERT statements independently, switch between one multi-row INSERT and one statement per row, and set the batch size. The parser is RFC-4180 friendly: quoted fields keep embedded commas and newlines, and doubled quotes are treated as escapes. Output updates live as you type, with one-click examples for users, products, orders, and employees, and a fully shareable URL. Everything runs in your browser.

How to Use

1Click an example CSV (Users, Products, Orders, or Employees) or paste your own CSV with a header row.
2Set the table name and pick a SQL dialect: PostgreSQL, MySQL, SQLite, or SQL Server.
3Choose a delimiter (or leave it on auto-detect) and confirm whether the first row is a header.
4Review the detected columns and override any inferred type from the preview table.
5Toggle CREATE TABLE, INSERT statements, and multi-row batching, then copy, download, or share the SQL.

Key Benefits

Infers integer, decimal, boolean, date, or VARCHAR/TEXT per column, with a per-column override
Dialect-aware output for PostgreSQL, MySQL, SQLite, and SQL Server
Robust CSV parsing: quoted fields with embedded commas and newlines, escaped quotes, auto-detected delimiter
Empty cells become NULL and single quotes are escaped automatically
Independent CREATE TABLE and INSERT toggles, plus multi-row batching with a configurable batch size
Live output, built-in example datasets, and a fully shareable URL
Runs entirely in your browser - no CSV ever leaves your machine

Common Use Cases

Importing a spreadsheet export into a database without writing schema by hand
Bootstrapping a table from a CSV during early prototyping
Generating seed data INSERTs from a sample dataset
Inspecting how columns would be typed before designing a real schema
Converting a CSV report into INSERT statements for a test database

Detected columns5 cols · 3 rows

Column	SQL type	Override
first_name	VARCHAR(50)
last_name	VARCHAR(50)
email	VARCHAR(50)
age	INT
city	VARCHAR(50)

SQL output

Copy, download, share, or pipe to another tool

CREATE TABLE "users" (
  "first_name" VARCHAR(50) NOT NULL,
  "last_name" VARCHAR(50) NOT NULL,
  "email" VARCHAR(50) NOT NULL,
  "age" INT NOT NULL,
  "city" VARCHAR(50) NOT NULL
);

INSERT INTO "users" ("first_name", "last_name", "email", "age", "city") VALUES
  ('John', 'Doe', '[email protected]', 30, 'New York'),
  ('Jane', 'Smith', '[email protected]', 25, 'Los Angeles'),
  ('Michael', 'Johnson', '[email protected]', 35, 'Chicago');

Target

Table name

SQL dialect

Examples:

Options

Delimiter

Rows per INSERT

First row is a header

Multi-row INSERT

Include CREATE TABLE

Include INSERT statements

CSV input

175 characters · paste CSV or load an example

Robust parsing

The parser is RFC-4180 friendly: quoted fields keep embedded commas and newlines, "" is treated as an escaped quote, and the delimiter is auto-detected (or pick comma, semicolon, tab, or pipe yourself).

Dialect-aware output

Types and quoting follow your dialect: PostgreSQL gets double quotes and TRUE/FALSE, MySQL gets backticks and TINYINT(1), SQL Server gets brackets and BIT. Empty cells become NULL, and you can override any inferred column type from the preview table.

Everything runs in your browser - no CSV ever leaves your machine.

Was this tool helpful?

Share Your Experience

Help others discover this tool!

Share on

CSV to SQL Converter

How to Use

Key Benefits

Common Use Cases

Target

Options

Robust parsing

Dialect-aware output

Was this tool helpful?

Share Your Experience

Related tools

Frequently Asked Questions - CSV to SQL Converter