Why Functions returning TABLE in PostgreSQL? - Purpose & Use Cases

The Big Idea

What if you could get fresh, filtered data instantly with just one simple call?

The Scenario

Imagine you have a big spreadsheet with sales data, and you want to find all sales above a certain amount. You try to do this by copying data into new sheets and manually filtering rows every time you need an update.

The Problem

This manual filtering is slow and tiring. You might make mistakes copying data, and every time new sales come in, you have to repeat the whole process. It's easy to lose track or miss important rows.

The Solution

Functions returning TABLE let you write a reusable query that acts like a mini-database inside your database. You just call the function with your filter, and it returns exactly the rows you want, fresh and correct every time.

Before vs After

✗ Before

SELECT * FROM sales WHERE amount > 1000; -- run manually each time

✓ After

CREATE FUNCTION get_big_sales(min_amount numeric) RETURNS TABLE(id int, amount numeric) AS $$
BEGIN
  RETURN QUERY SELECT id, amount FROM sales WHERE amount > min_amount;
END;
$$ LANGUAGE plpgsql;

SELECT * FROM get_big_sales(1000);

What It Enables

You can build powerful, reusable queries that return tables, making your data work for you automatically and reliably.

Real Life Example

A store manager can quickly get a list of all customers who spent more than $500 last month by calling a function, without digging through raw data or spreadsheets.

Key Takeaways

Manual filtering is slow and error-prone.

Functions returning TABLE automate and simplify data retrieval.

They make queries reusable and easy to maintain.

Practice

1. What does a PostgreSQL function declared with RETURNS TABLE do?

easy

A. Returns no result, only performs actions

B. Returns only a single scalar value

C. Returns multiple rows and columns as a table result

D. Returns a JSON object

Solution

1. Step 1: Understand the purpose of RETURNS TABLE

   The RETURNS TABLE clause defines that the function will return a set of rows with specified columns, like a table.

2. Step 2: Compare with other return types

   Unlike scalar returns or void, RETURNS TABLE returns multiple rows and columns.

3. Final Answer:

   Returns multiple rows and columns as a table result -> Option C

4. Quick Check:

   RETURNS TABLE means multiple rows/columns [OK]

Hint: RETURNS TABLE means function outputs rows and columns [OK]

Common Mistakes:

• Thinking RETURNS TABLE returns a single value
• Confusing RETURNS TABLE with RETURNS VOID
• Assuming it returns JSON automatically

2. Which of the following is the correct syntax to declare a PostgreSQL function returning a table with columns id INT and name TEXT?

easy

A. CREATE FUNCTION f() RETURNS TABLE(id TEXT, name INT) AS $$ BEGIN RETURN QUERY SELECT 1, 'a'; END; $$ LANGUAGE plpgsql;

B. CREATE FUNCTION f() RETURNS SETOF RECORD AS $$ BEGIN RETURN QUERY SELECT 1, 'a'; END; $$ LANGUAGE plpgsql;

C. CREATE FUNCTION f() RETURNS INT AS $$ BEGIN RETURN 1; END; $$ LANGUAGE plpgsql;

D. CREATE FUNCTION f() RETURNS TABLE(id INT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT 1, 'a'; END; $$ LANGUAGE plpgsql;

Solution

1. Step 1: Check RETURNS TABLE syntax

   CREATE FUNCTION f() RETURNS TABLE(id INT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT 1, 'a'; END; $$ LANGUAGE plpgsql; correctly declares RETURNS TABLE(id INT, name TEXT) matching the column names and types.

2. Step 2: Verify RETURN QUERY usage

   CREATE FUNCTION f() RETURNS TABLE(id INT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT 1, 'a'; END; $$ LANGUAGE plpgsql; uses RETURN QUERY SELECT 1, 'a'; which returns rows matching the table structure.

3. Final Answer:

   Correct RETURNS TABLE syntax and return statement -> Option D

4. Quick Check:

   RETURNS TABLE with matching columns and RETURN QUERY [OK]

Hint: RETURNS TABLE needs column names/types and RETURN QUERY [OK]

Common Mistakes:

• Using RETURNS SETOF RECORD without column definition
• Swapping column types in RETURNS TABLE
• Returning scalar instead of query

3. Given this function:

CREATE FUNCTION get_numbers() RETURNS TABLE(num INT) AS $$ BEGIN RETURN QUERY SELECT generate_series(1,3); END; $$ LANGUAGE plpgsql;

What will be the output of SELECT * FROM get_numbers();?

medium

A. 3 rows with values 1, 2, 3

B. 1 row with value 3

C. Error: function does not return a table

D. Empty result set

Solution

1. Step 1: Understand generate_series(1,3)

   This function generates rows with values 1, 2, and 3.

2. Step 2: RETURN QUERY returns all rows

   The function returns all rows from generate_series(1,3) as a table with column num.

3. Final Answer:

   3 rows with values 1, 2, 3 -> Option A

4. Quick Check:

   generate_series(1,3) returns 3 rows [OK]

Hint: generate_series returns multiple rows, so function returns them all [OK]

Common Mistakes:

• Thinking it returns a single row
• Confusing RETURNS TABLE with scalar return
• Expecting an error due to missing RETURNS SETOF

4. Identify the error in this function:

CREATE FUNCTION get_data() RETURNS TABLE(id INT, val TEXT) AS $$ BEGIN RETURN SELECT 1, 'a'; END; $$ LANGUAGE plpgsql;

medium

A. Missing RETURN QUERY before SELECT

B. Wrong column types in RETURNS TABLE

C. Function must return VOID

D. Missing LANGUAGE declaration

Solution

1. Step 1: Check RETURN statement in RETURNS TABLE function

   In PL/pgSQL, to return rows from a query, use RETURN QUERY, not just RETURN.

2. Step 2: Identify missing RETURN QUERY

   The function uses RETURN SELECT which is invalid syntax; it should be RETURN QUERY SELECT.

3. Final Answer:

   Missing RETURN QUERY before SELECT -> Option A

4. Quick Check:

   Use RETURN QUERY to return rows in RETURNS TABLE [OK]

Hint: Use RETURN QUERY, not RETURN, to return table rows [OK]

Common Mistakes:

• Using RETURN instead of RETURN QUERY for table results
• Omitting LANGUAGE plpgsql
• Incorrect RETURNS TABLE column types

5. You want to create a function that returns a table of user IDs and their uppercase names from a users table with columns id INT and name TEXT. Which function definition correctly achieves this?

hard

A. CREATE FUNCTION get_upper_users() RETURNS TABLE(id INT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT id, LOWER(name) FROM users; END; $$ LANGUAGE plpgsql;

B. CREATE FUNCTION get_upper_users() RETURNS TABLE(id INT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT id, UPPER(name) FROM users; END; $$ LANGUAGE plpgsql;

C. CREATE FUNCTION get_upper_users() RETURNS SETOF record AS $$ BEGIN RETURN QUERY SELECT id, UPPER(name) FROM users; END; $$ LANGUAGE plpgsql;

D. CREATE FUNCTION get_upper_users() RETURNS TABLE(id TEXT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT id, UPPER(name) FROM users; END; $$ LANGUAGE plpgsql;

Solution

1. Step 1: Match RETURNS TABLE columns with SELECT output

   The function returns id INT and name TEXT, so the SELECT must output these types. CREATE FUNCTION get_upper_users() RETURNS TABLE(id INT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT id, UPPER(name) FROM users; END; $$ LANGUAGE plpgsql; matches this.

2. Step 2: Check the transformation applied

   CREATE FUNCTION get_upper_users() RETURNS TABLE(id INT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT id, UPPER(name) FROM users; END; $$ LANGUAGE plpgsql; uses UPPER(name) to convert names to uppercase as required.

3. Step 3: Verify other options

   CREATE FUNCTION get_upper_users() RETURNS TABLE(id INT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT id, LOWER(name) FROM users; END; $$ LANGUAGE plpgsql; uses LOWER instead of UPPER, CREATE FUNCTION get_upper_users() RETURNS SETOF record AS $$ BEGIN RETURN QUERY SELECT id, UPPER(name) FROM users; END; $$ LANGUAGE plpgsql; returns SETOF record (no column definition), CREATE FUNCTION get_upper_users() RETURNS TABLE(id TEXT, name TEXT) AS $$ BEGIN RETURN QUERY SELECT id, UPPER(name) FROM users; END; $$ LANGUAGE plpgsql; mismatches id type (TEXT instead of INT).

4. Final Answer:

   Correctly returns id and uppercase name as a table -> Option B

5. Quick Check:

   RETURNS TABLE with matching columns and UPPER(name) [OK]

Hint: Match RETURNS TABLE columns and use RETURN QUERY with correct SELECT [OK]

Common Mistakes:

• Using LOWER instead of UPPER
• Mismatching column types in RETURNS TABLE
• Using SETOF record instead of RETURNS TABLE