SQL INSERT to CSV Converter

How to use

1. 1Paste one or more SQL INSERT statements into the input, or press Load sample to see a worked example with multi-row VALUES and escaped quotes.
2. 2Pick an output format: CSV or TSV for spreadsheets, JSON for an array of objects, or Markdown for a documentation table.
3. 3Choose what NULL should become: an empty cell, the word NULL, or a custom token like \N. Toggle the header row for CSV and TSV if you do not want column names.
4. 4Leave Merge all statements on to combine every INSERT into one table, or turn it off to export only the first table when a paste mixes several.
5. 5Read the extracted rows on the right. The status badges show how many rows, columns, and tables were parsed, and any skipped statements are listed as warnings.
6. 6Press Copy output or Download to save the result. Parsing stays in your browser and nothing is uploaded.

About this tool

SQL INSERT to CSV Converter is the reverse of generating INSERT statements: it reads the INSERT statements you find in a database dump, a seed script, or a migration file and pulls the actual rows back out as CSV, TSV, JSON, or a Markdown table. People reach for it when the data they need is trapped inside a .sql file and the destination is a spreadsheet, a notebook, a JSON fixture, or a quick documentation table, and copying the values out by hand would be slow and error prone. The parser is a small SQL scanner rather than a naive regular expression, so it copes with the things that break simpler tools. It understands all three string-literal escaping styles seen in real exports: doubled single quotes such as 'it''s', backslash escapes such as 'it\'s' and '\n' from MySQL dumps, and PostgreSQL dollar-quoted strings written as $$ text $$. It reads identifiers quoted with backticks, double quotes, or square brackets as well as bare table and column names, and it strips line comments, hash comments, and block comments before parsing. It handles both single-row inserts and multi-row VALUES lists, several statements in one paste, INSERT IGNORE and REPLACE INTO, schema-qualified table names, and trailing clauses like ON CONFLICT or ON DUPLICATE KEY without choking on them. Statements that have no explicit column list still work; the columns are named col1, col2, and so on in order. Each value is classified as it is read. NULL is recognized and mapped to whatever you choose, an empty cell by default, the literal word NULL, or a custom token such as \N for re-importing. Numbers and TRUE or FALSE are detected so the JSON output is correctly typed instead of wrapping everything in strings, and string values are unescaped to their real text, with an optional trim. When a paste contains several INSERT statements you can merge them all into one result, with differing column sets unioned and gaps filled, or export only the first table, and the tool tells you how many rows, columns, and distinct tables it found. CSV and TSV output is quoted to the usual rules so commas, quotes, and newlines inside a value survive a round trip into Excel or Google Sheets, JSON is emitted as an array of objects keyed by column name, and the Markdown table escapes pipes and line breaks so it renders cleanly in a README. Everything runs in your browser. The SQL you paste, including data from private databases, is never uploaded, logged, or stored.

Free to use. Works in your browser. No signup, no login.