How do I reference my CSV in the query?

As a single ? placeholder, not a named table. For example: SELECT * FROM ? WHERE score > 80. Behind the scenes, your parsed CSV rows are passed in as the query's one and only parameter, and alasql binds them to that ?.

Does any of my data leave the browser?

No. The CSV parsing and the SQL query itself both run entirely client-side via the alasql JavaScript SQL engine, nothing is sent to a server.

What SQL features are supported?

Whatever alasql supports when querying an in-memory array of objects: SELECT, WHERE, GROUP BY, ORDER BY, JOIN across multiple loaded tables, aggregate functions like COUNT and SUM, and more. It's a real SQL engine, not a toy subset.

What happens if my query has a typo or references a column that doesn't exist?

The query throws an error, which is caught and shown as a friendly "Query failed" message rather than crashing the page, so you can fix the SQL and try again.

Query CSV with SQL - Run SQL Queries on a CSV Online

Overview

Introduction

Sometimes filtering or sorting a CSV by hand isn't enough, you want to GROUP BY a column, JOIN two datasets, or run an aggregate, the kind of question SQL answers naturally.

Query CSV with SQL lets you do exactly that: paste a CSV, write a real SQL query against it, and see the results as a table, all without installing a database.

What Is Query CSV with SQL?

An in-browser SQL query tool for CSV data, built on alasql, a JavaScript SQL engine that can query plain in-memory arrays of objects as if they were database tables.

Rather than naming your CSV as a table you have to CREATE first, it's passed to the query as a single positional parameter, referenced in your SQL text with a ? placeholder, similar to a parameterized query in any SQL client.

How Query CSV with SQL Works

The pasted CSV is parsed with this site's shared grid parser, then converted into an array of plain row objects keyed by the header row, exactly the shape alasql expects for querying an in-memory table.

Your SQL text and that row array are passed to alasql(sql, [rows]), which binds the array to the ? placeholder in your query and returns the result as an array of row objects, rendered below as a table.

When To Use Query CSV with SQL

Use this whenever a CSV question is naturally a SQL question, top N rows by some column, grouping and counting, filtering on multiple conditions, joining two small datasets together.

It's also a fast way to sanity-check a SQL query you're planning to run somewhere else, against a small CSV sample, before pointing it at a real database.

Often used alongside CSV Editor and CSV Viewer.

Features

Advantages

Real SQL, not a simplified filter builder, GROUP BY, JOIN, ORDER BY, and aggregate functions all work.
No setup: no database to install, no table schema to define, just paste and query.
Runs entirely in the browser, so nothing about your CSV data is sent anywhere.

Limitations

Every value from the CSV starts out as a string, so numeric comparisons in WHERE clauses may need an explicit cast depending on how alasql infers types for your query.
Very large CSVs held as an in-memory JavaScript array can be slower than a real database for complex queries.

Examples

Filtering rows with WHERE

Input

id,name,score
1,Ada,95
2,Grace,72
3,Alan,88

Query: SELECT * FROM ? WHERE score > 80

Output

[{"id":"1","name":"Ada","score":"95"},{"id":"3","name":"Alan","score":"88"}]

The ? placeholder is bound to the parsed CSV rows, and the WHERE clause filters to just the rows where score is greater than 80.

Best Practices & Notes

Best Practices

Start with SELECT * FROM ? to confirm your CSV parsed the way you expect before adding WHERE or GROUP BY clauses.
Use column aliases (SELECT name AS "Full Name") to make the results table's headers more readable.
Remember every value from `parseCsvGrid` starts out as a string, wrap a column in a numeric cast if you need to sort or compare it numerically rather than lexically.

Developer Notes

queryCsvWithSql in query-csv-with-sql.ts parses the CSV with parseCsvGrid, converts the header-plus-rows grid into Record<string, string>[] objects, then calls alasql(sql, [rows]) inside a try/catch (alasql throws synchronously on malformed SQL). The result columns are read from Object.keys(rows[0]) when at least one row comes back, otherwise an empty column list is treated as a valid zero-row result rather than an error.

Query CSV with SQL Use Cases

Filtering and sorting a CSV export with real WHERE and ORDER BY clauses instead of manual column tricks
Grouping and aggregating CSV data, counts, sums, averages, per category with GROUP BY
Sanity-checking a SQL query against a small CSV sample before running it against a production database

Common Mistakes

Referencing the CSV by a table name instead of ?, alasql needs the positional placeholder here since the data is passed as a parameter, not a pre-declared table.
Assuming numeric columns are already numbers, every CSV cell starts out as a string, so a WHERE clause comparing against a number may need an explicit CAST or conversion depending on your query.

Tips

Use CSV Viewer first if you just want to sort and eyeball the data, save this tool for when you actually need a WHERE, GROUP BY, or JOIN.
Pair with CSV Editor if a query reveals rows you want to fix directly, rather than just filter out.

Query CSV with SQL