# Database datapools

```{note}
The integrations the team actively uses and supports are listed on the
{doc}`front page <../index>`: Kafka, HTTP and InfluxDB. Database datapools
are not currently one of them, but it remains available, documented, and
open to contributions — if you rely on it or want to improve it, it's a
good candidate.
```

Install: `pip install mite[db]` (no functional change today — every mite
install already includes it; this will matter starting in mite 3.0).

## Overview

`mite_db_datapool` provides database-backed data pools for Mite testing scenarios. It allows you to load test data directly from databases and use it in your load testing journeys.

## Why SQLAlchemy?

We chose **SQLAlchemy** over **pyodbc** for several reasons:

- **Database Agnostic**: Works with SQLite, PostgreSQL, MySQL, SQL Server, Oracle, etc.
- **Connection Pooling**: Built-in connection management and pooling
- **Simpler API**: Clean, Pythonic interface for database operations
- **Better Error Handling**: More informative error messages
- **No Driver Dependencies**: SQLAlchemy handles driver selection automatically
- **URL-based Configuration**: Easy connection string format

## Files

### `db_to_pool.py`

Core implementation containing:

- `DBIterableDataPool`: Load data once, consume sequentially (burned data)
- `DBRecyclableIterableDataPool`: Load data once, reuse infinitely
- Memory protection with `max_size` parameter to prevent OOM with large datasets

### `mite_db_example.py`

Working example showing:

- SQLite database creation with test data
- DataPool setup and configuration
- Integration with Mite journey functions
- Basic load testing scenario

## Testing

Run the example:

```bash
# From mite directory
python mite_db_datapool/mite_db_example.py
```

Run unit tests:

```bash
pytest test/test_db_to_pool.py -v
```