---
title: GDPR-compliant data deletion
meta:
    description: Learn how to delete user data in Tinybird to maintain GDPR compliance using various methods and best practices.
---

# GDPR-compliant data deletion in Tinybird

Tinybird provides several methods for deleting user data to maintain GDPR compliance. The best approach depends on your data volume and deletion frequency.

## Methods for data deletion

### Use the CLI

For smaller datasets or infrequent deletions, you can use the [`tb datasource delete`](/forward/dev-reference/commands/tb-datasource) command:

```bash
tb datasource delete [datasource_name] --sql-condition "user_id = 'user_to_delete'"
```

This method is suitable for datasets with up to a few million rows and deletion frequencies of up to 1000 operations per week.

### Use the Delete API Endpoint

For more frequent deletions, you can use the [Delete API Endpoint](/api-reference/datasource-api#post--v0-datasources-(.+)-delete):

```
POST /v0/datasources/(.+)/delete
```

This method is more efficient for regular deletion operations.

### Implement a deletion queue

For large-scale applications with frequent deletion requests:

- Create a `delete_user` table to queue deletion requests.
- Set up a scheduled job, such as a weekly job, to process the queue and delete user data from all relevant Data Sources.
- Use partitioning on large Data Sources to optimize deletion operations.

## Best practices

- Ensure deletions cascade to all relevant Data Sources and Materialized Views.
- For Data Sources with millions of rows, consider partitioning to improve deletion performance.
- Monitor the performance impact of deletion operations and adjust your strategy as your data grows.
- Implement TTL on Data Sources where appropriate to automatically remove old data.

## Considerations

- Deleting data using `ALTER TABLE ... DELETE WHERE` can have performance implications on large datasets.
- Balance between immediate deletion for GDPR compliance and system performance.
- Regularly review and optimize your deletion strategy as your data volume grows.