Connecting Python Applications to a Streaming Database

Build a real-time notification system where SQL materialized views detect events and trigger alerts. A streaming database continuously evaluates notification rules against incoming data.

Architecture

Events (Kafka/CDC) → RisingWave → Notification View → App polls or webhook

Implementation

-- Define notification rules as a materialized view
CREATE MATERIALIZED VIEW notifications AS
SELECT user_id, 'large_order' as notification_type,
  order_id, amount, order_time
FROM orders WHERE amount > 1000 AND order_time > NOW() - INTERVAL '5 minutes'

UNION ALL

SELECT user_id, 'rapid_orders' as notification_type,
  NULL as order_id, COUNT(*)::DECIMAL as amount, MAX(order_time)
FROM orders WHERE order_time > NOW() - INTERVAL '10 minutes'
GROUP BY user_id HAVING COUNT(*) > 5;

Delivery

# Poll the notification view
cursor.execute("SELECT * FROM notifications WHERE user_id = %s", (user_id,))
for notification in cursor.fetchall():
    send_push_notification(notification)

Or use RisingWave's webhook sink for push-based delivery.

Frequently Asked Questions

Can RisingWave push notifications directly?

RisingWave can sink to Kafka (for async processing) or use webhook sinks for HTTP-based push delivery. For pull-based, applications query the notification view via PostgreSQL.