Table Converter (Markdown, LaTeX, HTML, CSV)

Continue in the tool

More pages for this tool

csv to sql insert create table

CSV to SQL: automatic CREATE TABLE and INSERT

Paste the CSV (comma or semicolon), pick the dialect and table name. The SQL tab hands you a CREATE TABLE with types chosen from your data and batched INSERTs, ready to run in the database client.

How types and values are decided

The column is read top to bottom and the type comes from the whole set of values. Integers only give INT, and BIGINT if any value goes past 2,147,483,647. Integers and decimals together give DECIMAL(p,s), with p equal to the largest count of integer digits plus the largest count of decimal places; 18.90 next to 2.50 produces DECIMAL(4,2). YYYY-MM-DD dates count only if they exist on the calendar, so 2023-02-30 is not a DATE. The values true and false give BOOLEAN. Any mixed column, or one with a leading zero such as a postal code, stays text.
The dialects differ in ways that matter when you run the script. MySQL uses backticks for identifiers and treats the backslash as an escape inside strings, so C:\temp has to come out as C:\\temp. PostgreSQL and SQLite use double quotes and do not escape the backslash. In all of them, a single quote inside text is doubled. SQLite has no boolean or date type, so it stores 1 and 0 and dates as TEXT. Empty cells in numeric, date or boolean columns become NULL, because an empty string would be rejected in those types.
Column names become safe identifiers: accent-free, lowercase and with underscores instead of spaces, so "Product Name" becomes product_name. Repeated names get a _2, _3 suffix, and a name that would start with a digit gets the c_ prefix. The batch size controls how many rows go into each INSERT; 100 is a good starting point because scripts with thousands of rows in a single statement can exceed the server packet limit. Numbers with a thousands separator and decimal comma, such as 1.018,90, are written as 1018.90.

Checked examples

Brazilian CSV to MySQL

Input: id;nome;preco;ativo;criado 1;Caneta;2,50;true;2026-03-01 2;Caderno;18,90;false;2026-03-02
Expected output: CREATE TABLE `produtos` ( `id` INT, `nome` VARCHAR(16), `preco` DECIMAL(4,2), `ativo` BOOLEAN, `criado` DATE ); INSERT INTO `produtos` (`id`, `nome`, `preco`, `ativo`, `criado`) VALUES (1, 'Caneta', 2.50, TRUE, '2026-03-01'), (2, 'Caderno', 18.90, FALSE, '2026-03-02');

Table named produtos, MySQL dialect. The decimal comma became a dot.

Leading-zero postal code on PostgreSQL

Input: cep,cidade 01001000,São Paulo 20040020,Rio de Janeiro
Expected output: CREATE TABLE "cidades" ( "cep" TEXT, "cidade" TEXT ); INSERT INTO "cidades" ("cep", "cidade") VALUES ('01001000', 'São Paulo'), ('20040020', 'Rio de Janeiro');

Table cidades, PostgreSQL dialect. The cep column stays TEXT to preserve the zero.

Quotes and backslash, INSERT only

Input: nome,obs Ana,"disse ""oi""" Bia,C:\temp
Expected output: INSERT INTO `notas` (`nome`, `obs`) VALUES ('Ana', 'disse "oi"'), ('Bia', 'C:\\temp');

Table notas, MySQL, no CREATE TABLE. The backslash is doubled.

Large integer becomes BIGINT

Input: qtd,valor 3000000000,1.018
Expected output: CREATE TABLE `my_table` ( `qtd` BIGINT, `valor` DECIMAL(4,3) ); INSERT INTO `my_table` (`qtd`, `valor`) VALUES (3000000000, 1.018);

Default options, MySQL, name my_table. 3,000,000,000 exceeds 2,147,483,647.

Full tool FAQ

How do I convert an Excel table to Markdown?

Select the cells in Excel, copy with Ctrl+C and paste into the input box: the content arrives tab-separated and is recognized right away. Keep the Markdown tab active and click Copy. Numeric columns come out right-aligned (---: in the hyphen row) and text columns left-aligned, and you can force another alignment in the editing toolbar. Formulas arrive as the displayed value, not as the formula.

How do I convert CSV to a LaTeX table?

Is semicolon-separated CSV with a decimal comma read correctly?

How do I write the | character inside a Markdown table cell?

Can I merge cells?

How does SQL pick the column types?

Is there a row or column limit?

Can I convert the other way, from Markdown, HTML, JSON or LaTeX to something else?

Is my data sent to a server?

Why do 007 or a ZIP code with a leading zero stay text in SQL and JSON?

Frequently asked questions

Are the inferred types final?

They are a suggestion based on the data you pasted. A column that today only holds values up to 99.99 gets DECIMAL(4,2); if future data is larger, adjust the type. Turn inference off to get text only.

Does this protect against SQL injection?

Text is escaped so the script runs correctly, but the goal is generating a data load you review, not building queries from user input. In applications, use parameterized queries.

How many rows per INSERT should I use?

Between 100 and 1,000 usually works well. Very high values create huge statements, and MySQL may refuse a packet larger than its max_allowed_packet.

Next steps

How types and values are decided

The column is read top to bottom and the type comes from the whole set of values. Integers only give INT, and BIGINT if any value goes past 2,147,483,647. Integers and decimals together give DECIMAL(p,s), with p equal to the largest count of integer digits plus the largest count of decimal places; 18.90 next to 2.50 produces DECIMAL(4,2). YYYY-MM-DD dates count only if they exist on the calendar, so 2023-02-30 is not a DATE. The values true and false give BOOLEAN. Any mixed column, or one with a leading zero such as a postal code, stays text.

The dialects differ in ways that matter when you run the script. MySQL uses backticks for identifiers and treats the backslash as an escape inside strings, so C:\temp has to come out as C:\\temp. PostgreSQL and SQLite use double quotes and do not escape the backslash. In all of them, a single quote inside text is doubled. SQLite has no boolean or date type, so it stores 1 and 0 and dates as TEXT. Empty cells in numeric, date or boolean columns become NULL, because an empty string would be rejected in those types.

Column names become safe identifiers: accent-free, lowercase and with underscores instead of spaces, so "Product Name" becomes product_name. Repeated names get a _2, _3 suffix, and a name that would start with a digit gets the c_ prefix. The batch size controls how many rows go into each INSERT; 100 is a good starting point because scripts with thousands of rows in a single statement can exceed the server packet limit. Numbers with a thousands separator and decimal comma, such as 1.018,90, are written as 1018.90.

Checked examples

Brazilian CSV to MySQL

Input: id;nome;preco;ativo;criado 1;Caneta;2,50;true;2026-03-01 2;Caderno;18,90;false;2026-03-02
Expected output: CREATE TABLE `produtos` ( `id` INT, `nome` VARCHAR(16), `preco` DECIMAL(4,2), `ativo` BOOLEAN, `criado` DATE ); INSERT INTO `produtos` (`id`, `nome`, `preco`, `ativo`, `criado`) VALUES (1, 'Caneta', 2.50, TRUE, '2026-03-01'), (2, 'Caderno', 18.90, FALSE, '2026-03-02');

Table named produtos, MySQL dialect. The decimal comma became a dot.

Leading-zero postal code on PostgreSQL

Input: cep,cidade 01001000,São Paulo 20040020,Rio de Janeiro
Expected output: CREATE TABLE "cidades" ( "cep" TEXT, "cidade" TEXT ); INSERT INTO "cidades" ("cep", "cidade") VALUES ('01001000', 'São Paulo'), ('20040020', 'Rio de Janeiro');

Table cidades, PostgreSQL dialect. The cep column stays TEXT to preserve the zero.

Quotes and backslash, INSERT only

Input: nome,obs Ana,"disse ""oi""" Bia,C:\temp
Expected output: INSERT INTO `notas` (`nome`, `obs`) VALUES ('Ana', 'disse "oi"'), ('Bia', 'C:\\temp');

Table notas, MySQL, no CREATE TABLE. The backslash is doubled.

Large integer becomes BIGINT

Input: qtd,valor 3000000000,1.018
Expected output: CREATE TABLE `my_table` ( `qtd` BIGINT, `valor` DECIMAL(4,3) ); INSERT INTO `my_table` (`qtd`, `valor`) VALUES (3000000000, 1.018);

Default options, MySQL, name my_table. 3,000,000,000 exceeds 2,147,483,647.

Full tool FAQ

How do I convert an Excel table to Markdown?

How do I convert CSV to a LaTeX table?

Is semicolon-separated CSV with a decimal comma read correctly?

How do I write the | character inside a Markdown table cell?

Can I merge cells?

How does SQL pick the column types?

Is there a row or column limit?

Can I convert the other way, from Markdown, HTML, JSON or LaTeX to something else?

Is my data sent to a server?

Why do 007 or a ZIP code with a leading zero stay text in SQL and JSON?

Frequently asked questions

Are the inferred types final?

Does this protect against SQL injection?

Text is escaped so the script runs correctly, but the goal is generating a data load you review, not building queries from user input. In applications, use parameterized queries.

How many rows per INSERT should I use?

Between 100 and 1,000 usually works well. Very high values create huge statements, and MySQL may refuse a packet larger than its max_allowed_packet.

Table Converter (Markdown, LaTeX, HTML, CSV)

Continue in the tool

More pages for this tool

CSV to SQL: automatic CREATE TABLE and INSERT

How types and values are decided

Checked examples

Brazilian CSV to MySQL

Leading-zero postal code on PostgreSQL

Quotes and backslash, INSERT only

Large integer becomes BIGINT

Full tool FAQ

Frequently asked questions

Are the inferred types final?

Does this protect against SQL injection?

How many rows per INSERT should I use?

Next steps

Table Converter (Markdown, LaTeX, HTML, CSV)

Cell editor

Output

Continue in the tool

More pages for this tool

CSV to SQL: automatic CREATE TABLE and INSERT

How types and values are decided

Checked examples

Brazilian CSV to MySQL

Leading-zero postal code on PostgreSQL

Quotes and backslash, INSERT only

Large integer becomes BIGINT

Full tool FAQ

Frequently asked questions

Are the inferred types final?

Does this protect against SQL injection?

How many rows per INSERT should I use?

Next steps

Cell editor

Output