Compare Datasets in Browser — Diff CSV, Parquet, JSON

Need to diff two CSV exports? Reconcile last month’s report against this month’s? Validate that a data migration didn’t drop rows? PondPilot makes dataset comparison straightforward with SQL.

Open Both Files, Write SQL

Open two files in the same PondPilot session and use SQL to find differences:

-- Rows in new export but not in old
SELECT * FROM new_export.csv
EXCEPT
SELECT * FROM old_export.csv;
-- Row counts by status in both
SELECT 'old' as source, status, COUNT(*) FROM old.csv GROUP BY status
UNION ALL
SELECT 'new' as source, status, COUNT(*) FROM new.csv GROUP BY status
ORDER BY status, source;

No diffing tool with weird UIs. Just SQL against your actual data.

Common Comparison Workflows

Data migration validation: Export before and after, query for missing or changed records.

Report reconciliation: Compare monthly exports to catch discrepancies early.

Schema drift detection: Use DESCRIBE on both files to spot column changes.

Deduplication: Find duplicate records across or within datasets using GROUP BY and HAVING.

Mix File Formats

Your old export is CSV and the new one is Parquet? No problem. PondPilot treats all supported formats as queryable tables. Join a CSV against a Parquet file in the same query.

All Local, All Private

Both files stay on your machine. PondPilot has no backend server — DuckDB runs as WebAssembly in your browser. This matters when you’re comparing production data or sensitive records.

Export the Diff

Found the differences? Export the query results to CSV and share with your team. Or keep iterating with more queries until you’ve fully understood what changed.

Get Started

Open PondPilot, drop both files in, and start comparing. Free, private, no account needed.