Indexing materialized views in PostgreSQL - Step-by-Step Execution

Concept Flow - Indexing materialized views

Create Materialized View

↓

Materialized View Created

↓

Create Index on Materialized View

↓

Index Created

↓

Query Materialized View

↓

Index Used to Speed Up Query

↓

Refresh Materialized View as Needed

First, create a materialized view, then create an index on it to speed up queries. Refresh the view when data changes.

Execution Sample

PostgreSQL

CREATE MATERIALIZED VIEW sales_summary AS
SELECT product_id, SUM(amount) AS total_sales
FROM sales
GROUP BY product_id;

CREATE INDEX idx_sales_summary_product ON sales_summary(product_id);

This code creates a materialized view summarizing sales by product and then creates an index on product_id to speed up queries.

Execution Table

Step	Action	Object Affected	Result	Notes
1	Execute CREATE MATERIALIZED VIEW	sales_summary	Materialized view created with aggregated data	Data snapshot stored
2	Execute CREATE INDEX	idx_sales_summary_product	Index created on sales_summary(product_id)	Speeds up queries filtering by product_id
3	Query materialized view filtering by product_id	sales_summary	Query uses index for faster lookup	Improved query performance
4	Refresh materialized view	sales_summary	Data updated to reflect latest sales	Index remains valid
5	Query again after refresh	sales_summary	Index used again for fast query	Consistent performance

💡 Execution stops after queries use the index and materialized view is refreshed as needed

Variable Tracker

Object	Initial State	After Step 1	After Step 2	After Step 4	Final State
sales_summary	Does not exist	Created with data snapshot	Unchanged	Refreshed with new data	Up-to-date data snapshot
idx_sales_summary_product	Does not exist	Does not exist	Created on product_id	Unchanged	Index available for queries

Key Moments - 3 Insights

Why do we create an index on a materialized view?

Creating an index on the materialized view (see step 2 in execution_table) helps speed up queries that filter or join on the indexed column, improving performance.

Does the index need to be recreated after refreshing the materialized view?

No, the index remains valid after refreshing the materialized view (see step 4 and 5). Refresh updates data but keeps the index intact.

What happens if we query the materialized view without an index?

Queries will scan the entire materialized view, which can be slower (compare step 3 with and without index). Indexes help avoid full scans.

Visual Quiz - 3 Questions

Look at the execution_table, at which step is the index created on the materialized view?

A Step 2

B Step 1

C Step 3

D Step 4

💡 Hint

Check the 'Action' column in execution_table for 'CREATE INDEX'

According to variable_tracker, what is the state of sales_summary after step 4?

A Does not exist

B Created with data snapshot

C Refreshed with new data

D Index created

💡 Hint

Look at the 'sales_summary' row under 'After Step 4' in variable_tracker

If we do not create an index on the materialized view, what is the likely effect on queries?

A Queries run faster

B Queries run slower due to full scans

C Materialized view cannot be queried

D Materialized view refresh fails

💡 Hint

Refer to key_moments about query performance without index

Concept Snapshot

CREATE MATERIALIZED VIEW stores query results physically.
CREATE INDEX on materialized view speeds up queries.
Refresh materialized view to update data.
Index stays valid after refresh.
Use indexes to avoid full scans and improve performance.

Full Transcript

This visual execution shows how to create a materialized view in PostgreSQL, then create an index on it to speed up queries. First, the materialized view is created with aggregated data. Next, an index is created on a column to help queries filter faster. When querying the materialized view, the index is used to improve performance. Refreshing the materialized view updates the data but keeps the index intact. This process helps maintain fast query speeds on large, precomputed datasets.

Practice

1. What is the main purpose of creating an index on a materialized view in PostgreSQL?

easy

A. To speed up query performance on the materialized view

B. To automatically refresh the materialized view

C. To store the materialized view data physically

D. To prevent the materialized view from being updated

Solution

1. Step 1: Understand what an index does

   An index helps the database find data faster by creating a quick lookup structure.

2. Step 2: Apply this to materialized views

   Since materialized views store query results physically, indexing them speeds up queries on that stored data.

3. Final Answer:

   To speed up query performance on the materialized view -> Option A

4. Quick Check:

   Indexing = Faster queries [OK]

Hint: Indexes speed up data retrieval on stored results [OK]

Common Mistakes:

• Thinking indexes refresh the view automatically
• Confusing storage with indexing
• Believing indexes prevent updates

2. Which of the following is the correct syntax to create an index on a materialized view named sales_summary on column region?

easy

A. CREATE INDEX idx_region ON TABLE sales_summary(region);

B. CREATE INDEX idx_region ON MATERIALIZED VIEW sales_summary(region);

C. CREATE INDEX idx_region ON sales_summary(region);

D. CREATE INDEX idx_region ON VIEW sales_summary(region);

Solution

1. Step 1: Recall index creation syntax

   Indexes are created with CREATE INDEX index_name ON table_name(column); syntax.

2. Step 2: Apply to materialized views

   Materialized views behave like tables for indexing, so no extra keywords like MATERIALIZED VIEW or VIEW are used.

3. Final Answer:

   CREATE INDEX idx_region ON sales_summary(region); -> Option C

4. Quick Check:

   Index syntax = CREATE INDEX name ON table(column) [OK]

Hint: Use table syntax for materialized views when creating indexes [OK]

Common Mistakes:

• Adding MATERIALIZED VIEW keyword in index creation
• Using VIEW keyword instead of table name
• Incorrect syntax with TABLE keyword

3. Given a materialized view mv_orders with an index on order_date, what will be the output of this query?

SELECT order_date FROM mv_orders WHERE order_date = '2023-01-01';

Assuming the materialized view is refreshed and contains data for that date.

medium

A. Returns all rows with order_date '2023-01-01' quickly using the index

B. Returns an error because indexes cannot be used on materialized views

C. Returns no rows because materialized views do not store data

D. Returns all rows but ignores the index, causing slow performance

Solution

1. Step 1: Understand index usage on materialized views

   Indexes on materialized views speed up queries by quickly locating matching rows.

2. Step 2: Analyze the query behavior

   The query filters by order_date, which is indexed, so it returns matching rows efficiently.

3. Final Answer:

   Returns all rows with order_date '2023-01-01' quickly using the index -> Option A

4. Quick Check:

   Indexed query = fast matching rows [OK]

Hint: Indexed columns speed up matching queries on materialized views [OK]

Common Mistakes:

• Thinking indexes don't work on materialized views
• Assuming materialized views are empty without refresh
• Believing query ignores indexes always

4. You created an index on a materialized view but notice queries are still slow. What is the most likely cause?

medium

A. Indexes do not work on materialized views in PostgreSQL

B. The materialized view has not been refreshed after data changes

C. The index was created on a non-existent column

D. Materialized views cannot have indexes on date columns

Solution

1. Step 1: Understand materialized view refresh

   Materialized views store data physically but need manual refresh to update data and indexes.

2. Step 2: Connect refresh to query speed

   If not refreshed, the view and its index are outdated, causing slow or incorrect query results.

3. Final Answer:

   The materialized view has not been refreshed after data changes -> Option B

4. Quick Check:

   Refresh needed for updated data and index [OK]

Hint: Always refresh materialized views after data changes [OK]

Common Mistakes:

• Believing indexes don't work on materialized views
• Creating index on wrong column silently
• Thinking date columns can't be indexed

5. You have a materialized view mv_sales refreshed daily. You want to speed up queries filtering by customer_id and sale_date. Which is the best indexing strategy?

hard

A. Do not create any index; rely on sequential scan

B. Create separate indexes on customer_id and sale_date individually

C. Create an index only on sale_date because it is a date column

D. Create a single composite index on (customer_id, sale_date)

Solution

1. Step 1: Understand composite vs single indexes

   A composite index on multiple columns speeds up queries filtering by both columns together.

2. Step 2: Apply to query filtering by customer_id and sale_date

   Since queries filter on both columns, a composite index is more efficient than separate indexes.

3. Final Answer:

   Create a single composite index on (customer_id, sale_date) -> Option D

4. Quick Check:

   Composite index = faster multi-column filters [OK]

Hint: Use composite index for multi-column filtering queries [OK]

Common Mistakes:

• Creating separate indexes causing less efficient queries
• Indexing only one column when filtering by two
• Skipping indexes and relying on slow scans