SQL diff checker

How to compare two SQL queries

1. Paste the original query, file, or definition into the Original box.
2. Paste the version you want to check into the Modified box. The diff appears right away and updates with every keystroke.
3. Read the highlights. Red lines with a minus sign exist only in the original, green lines with a plus sign exist only in the modified version, and plain lines are identical on both sides.
4. If everything looks changed, the two versions are probably formatted differently. Run both through the SQL formatter and compare again.
5. Use the line numbers on each side to point a teammate to the exact change.

Example: compare two versions of a query

Say a revenue report originally counted every paid order. The new version limits the date range and returns only the top 20 customers. Pasting both versions produces this diff:

The new date filter and the LIMIT are easy to see. Note that ORDER BY revenue DESC; is also marked as removed. The semicolon moved to the new last line, and a line diff treats that as a changed line. That is expected behavior, and it is why the diff stays reliable on long queries: nothing is hidden or summarized.

Example: compare two migration files

Migration reviews are where a diff saves the most time. Here the original CREATE TABLE is compared with a revised version that adds a foreign key, widens a numeric column, and adds a paid_at column:

The diff surfaces the new foreign key, the wider numeric type, and the new column. The status line appears as removed and added even though only a trailing comma changed, because adding a column forces a comma onto the line above it. Knowing that pattern makes migration diffs quick to read. For whole-database comparisons across many tables, use the database schema diff, which reports added, removed, and changed tables and columns instead of raw lines.

Example: two queries that only differ in formatting

A line diff compares lines, so a query written on one line and the same query written across several lines look completely different. Compare these two:

-- Original
select id,name from users where active=true

-- Modified
SELECT
  id,
  name
FROM users
WHERE active = TRUE

Every line is flagged, although the two queries are the same. Format both with the same settings in the SQL formatter first, and the diff shrinks to the changes that matter. This is also the best way to verify that an auto-formatter or refactor did not change query logic.

Compare what two queries return

A text diff shows how two queries differ, not whether their results differ. When you need to know if a rewrite still returns the same rows, run both against your database and subtract one result from the other in each direction:

(SELECT id, email FROM users_v1)
EXCEPT
(SELECT id, email FROM users_v2);

Then swap the two queries and run it again. If both runs return zero rows, the results match. EXCEPT is available in PostgreSQL, SQL Server, SQLite, and MySQL 8.0.31 or later, BigQuery requires EXCEPT DISTINCT, and Oracle calls it MINUS.

Compare stored procedures, functions, and views

To compare database objects such as staging against production, extract each definition as text and paste both into the tool:

• PostgreSQL: pg_get_functiondef('my_function'::regproc) and pg_get_viewdef('my_view'::regclass)
• MySQL: SHOW CREATE PROCEDURE my_procedure and SHOW CREATE VIEW my_view
• SQL Server: OBJECT_DEFINITION(OBJECT_ID('dbo.my_procedure')) or sp_helptext

When to use a SQL diff tool

• Reviewing query changes in pull requests before merging
• Auditing database migration files by comparing the old and new versions of a CREATE TABLE, view, or stored procedure
• Debugging regressions by pasting the last working version of a query next to the current broken one
• Documenting how a query or schema evolved over time
• Verifying that a formatter or refactor did not change query logic

How the diff is computed

The tool uses a standard line-by-line diffing algorithm, the same approach as git diff. Lines present only in the original are marked as removed, lines present only in the modified version are marked as added, and matching lines appear unchanged with line numbers for both sides. Whitespace counts, so trailing spaces and tabs are differences too.

Stop writing SQL manually for dashboards

If you find yourself iterating on SQL queries to build reports and charts, Basedash can eliminate that loop entirely. Describe the metric or visualization you need in plain English, and Basedash generates the SQL, runs it against your database, and renders the result, with no manual query writing or diffing needed. It connects to PostgreSQL, MySQL, BigQuery, Snowflake, and 750+ other data sources.