การใช้มุมมองแบบจัดเก็บเพื่อเพิ่มประสิทธิภาพ

Welcome to the lesson on Materialized Views! These are powerful tools in PostgreSQL for speeding up complex queries.

Think of a Materialized View (MV) as a pre-computed result set of a query that is stored on disk. Unlike a regular view, which runs its query every time you access it, an MV holds the actual data.

Materialized Views are incredibly useful for performance, especially when dealing with:

• Complex Joins: Queries involving many tables.
• Aggregations: Calculations like sums, averages, or counts over large datasets.
• Reporting & Analytics: Dashboards and reports that frequently query the same complex data.

By pre-calculating and storing these results, MVs can drastically reduce query execution time for repeated requests.

You create a Materialized View using the CREATE MATERIALIZED VIEW statement, followed by a SELECT query. The query defines the data that will be stored in your MV.

Let's create a simple Materialized View to see the average price per category from a products table.

Try running this example. It sets up a small products table, inserts some data, and then creates a Materialized View named category_avg_price based on that data.

A key difference from regular views is that Materialized Views do not update automatically when the underlying data changes. You need to explicitly refresh them.

Use the REFRESH MATERIALIZED VIEW command to update the data in an MV. For large MVs, you can add CONCURRENTLY to allow other queries to access the view while it's refreshing.

Let's add a new product to our products table and then refresh our category_avg_price Materialized View. Notice how the MV shows the old data until it's refreshed.

It's important to understand the core differences between Materialized Views and standard views:

• Standard Views: Are essentially stored queries. They don't store data themselves; they execute their defining query every time they are accessed.
• Materialized Views: Store the result of their defining query as actual data on disk. This makes reading from them much faster, but they require manual (or scheduled) refreshing.

Choose MVs when performance is critical for complex, static-ish data.

Materialized Views are ideal for:

• Batch Reporting: Generating daily, weekly, or monthly reports that don't need real-time data.
• Data Warehousing: Pre-aggregating data for faster analytical queries.
• External Dashboards: Providing quick access to complex metrics for BI tools.
• Static Data: When the underlying tables don't change very frequently, minimizing refresh overhead.

Avoid MVs for highly transactional, real-time data that needs immediate updates.

Consider a Materialized View named daily_sales_summary that aggregates sales data from a transactions table.

If new transactions are added to the transactions table, what must you do for daily_sales_summary to reflect these new sales?

In this lesson, you've learned about Materialized Views in PostgreSQL.

• They are pre-computed query results stored on disk, offering significant performance gains for complex, repetitive queries.
• You create them with CREATE MATERIALIZED VIEW.
• They require explicit refreshing using REFRESH MATERIALIZED VIEW to update their data.
• They are perfect for reporting, analytics, and data warehousing where real-time updates aren't critical.

Keep practicing with MVs to master this powerful optimization technique!