Cursor declaration and usage in PostgreSQL

Introduction

Cursors let you look at query results one row at a time. This helps when you want to handle data step-by-step instead of all at once.

When you want to process large query results without loading everything into memory.

When you need to perform operations on each row individually.

When you want to fetch rows in small chunks for better control.

When you want to loop through query results inside a stored procedure or function.

When you want to avoid locking the whole table by processing rows gradually.

Syntax

PostgreSQL

DECLARE cursor_name CURSOR FOR select_statement;

OPEN cursor_name;

FETCH cursor_name INTO target_variable;

CLOSE cursor_name;

You declare a cursor with a name and a SELECT query.

Use OPEN to start the cursor, FETCH to get rows one by one, and CLOSE when done.

Examples

This declares a cursor named emp_cursor to select employee IDs and names.

PostgreSQL

DECLARE emp_cursor CURSOR FOR SELECT id, name FROM employees;

This opens the cursor and fetches the first row into variables emp_id and emp_name.

PostgreSQL

OPEN emp_cursor;
FETCH emp_cursor INTO emp_id, emp_name;

This closes the cursor to release resources.

PostgreSQL

CLOSE emp_cursor;

Sample Program

This anonymous code block declares a cursor to select employee IDs and names ordered by ID. It opens the cursor, fetches each row one by one, prints the employee details, and closes the cursor.

PostgreSQL

DO $$
DECLARE
  emp_cursor CURSOR FOR SELECT id, name FROM employees ORDER BY id;
  emp_id INT;
  emp_name TEXT;
BEGIN
  OPEN emp_cursor;
  LOOP
    FETCH emp_cursor INTO emp_id, emp_name;
    EXIT WHEN NOT FOUND;
    RAISE NOTICE 'Employee ID: %, Name: %', emp_id, emp_name;
  END LOOP;
  CLOSE emp_cursor;
END $$;

Important Notes

Cursors are useful for row-by-row processing but can be slower than set-based queries.

Always close cursors to free resources.

Use EXIT WHEN NOT FOUND; to stop fetching when no more rows are available.

Summary

Cursors let you handle query results one row at a time.

Declare, open, fetch rows, and close the cursor in that order.

Use cursors when you need step-by-step control over data processing.

Practice

1. What is the primary purpose of declaring a cursor in PostgreSQL?

easy

A. To speed up bulk inserts

B. To process query results one row at a time

C. To create a new table in the database

D. To backup the database automatically

Solution

1. Step 1: Understand what a cursor does

   A cursor allows you to handle query results row by row instead of all at once.

2. Step 2: Compare options with cursor purpose

   Only To process query results one row at a time describes this behavior; others describe unrelated tasks.

3. Final Answer:

   To process query results one row at a time -> Option B

4. Quick Check:

   Cursor purpose = process rows one by one [OK]

Hint: Cursors handle rows stepwise, not bulk operations [OK]

Common Mistakes:

• Confusing cursors with table creation
• Thinking cursors speed up inserts
• Assuming cursors automate backups

2. Which of the following is the correct syntax to declare a cursor named cur_emp for selecting all rows from employees table?

easy

A. CREATE CURSOR cur_emp AS SELECT * FROM employees;

B. OPEN cur_emp CURSOR SELECT * FROM employees;

C. DECLARE cur_emp CURSOR FOR SELECT * FROM employees;

D. FETCH cur_emp FROM employees;

Solution

1. Step 1: Recall cursor declaration syntax

   In PostgreSQL, cursors are declared with DECLARE cursor_name CURSOR FOR query.

2. Step 2: Match syntax with options

   DECLARE cur_emp CURSOR FOR SELECT * FROM employees; matches this exactly; others use incorrect keywords or order.

3. Final Answer:

   DECLARE cur_emp CURSOR FOR SELECT * FROM employees; -> Option C

4. Quick Check:

   DECLARE + CURSOR + FOR + query = correct syntax [OK]

Hint: Use DECLARE ... CURSOR FOR ... to declare [OK]

Common Mistakes:

• Using OPEN instead of DECLARE for declaration
• Confusing FETCH with DECLARE
• Using CREATE CURSOR which is invalid syntax

3. Given the following code snippet, what will be the output after fetching from the cursor?

DECLARE cur_emp CURSOR FOR SELECT id FROM employees ORDER BY id LIMIT 3;
OPEN cur_emp;
FETCH NEXT FROM cur_emp;
FETCH NEXT FROM cur_emp;

medium

A. First two employee ids in ascending order

B. All employee ids from the table

C. Syntax error due to missing CLOSE statement

D. Empty result because cursor is not opened

Solution

1. Step 1: Understand cursor declaration and fetch

   The cursor selects 3 employee ids ordered ascending. FETCH NEXT retrieves one row each time.

2. Step 2: Analyze fetch calls

   Two FETCH NEXT calls return the first two rows from the cursor result.

3. Final Answer:

   First two employee ids in ascending order -> Option A

4. Quick Check:

   Two FETCH NEXT = two rows fetched [OK]

Hint: Each FETCH returns one row in cursor order [OK]

Common Mistakes:

• Assuming FETCH returns all rows at once
• Thinking missing CLOSE causes syntax error
• Believing cursor must be closed before fetching

4. Identify the error in the following cursor usage:

DECLARE cur_dept CURSOR FOR SELECT name FROM departments;
FETCH NEXT FROM cur_dept;
OPEN cur_dept;
CLOSE cur_dept;

medium

A. Cursor declaration syntax is incorrect

B. Cursor is declared after fetching

C. Cursor is closed before declaration

D. Cursor is fetched before it is opened

Solution

1. Step 1: Check the order of cursor operations

   Cursors must be declared, then opened, then fetched, then closed.

2. Step 2: Identify incorrect sequence

   Here, FETCH is called before OPEN, which is invalid.

3. Final Answer:

   Cursor is fetched before it is opened -> Option D

4. Quick Check:

   OPEN must precede FETCH [OK]

Hint: Always OPEN cursor before FETCH [OK]

Common Mistakes:

• Fetching before opening cursor
• Closing cursor before opening
• Misordering declaration and fetch

5. You want to process all rows from orders table one by one using a cursor in a PL/pgSQL function. Which sequence of statements correctly implements this?

hard

A. DECLARE cur_orders CURSOR FOR SELECT * FROM orders; OPEN cur_orders; LOOP FETCH cur_orders INTO rec; EXIT WHEN NOT FOUND; -- process rec END LOOP; CLOSE cur_orders;

B. OPEN cur_orders; DECLARE cur_orders CURSOR FOR SELECT * FROM orders; FETCH cur_orders INTO rec; CLOSE cur_orders;

C. DECLARE cur_orders CURSOR FOR SELECT * FROM orders; FETCH cur_orders INTO rec; OPEN cur_orders; CLOSE cur_orders;

D. DECLARE cur_orders CURSOR FOR SELECT * FROM orders; OPEN cur_orders; FETCH ALL FROM cur_orders INTO rec; CLOSE cur_orders;

Solution

1. Step 1: Understand correct cursor usage in PL/pgSQL

   Declare cursor, open it, then loop fetching rows until no more rows, then close cursor.

2. Step 2: Analyze each option

   DECLARE cur_orders CURSOR FOR SELECT * FROM orders; OPEN cur_orders; LOOP FETCH cur_orders INTO rec; EXIT WHEN NOT FOUND; -- process rec END LOOP; CLOSE cur_orders; follows correct order and uses LOOP with EXIT WHEN NOT FOUND to process all rows. Others have wrong order or invalid FETCH ALL.

3. Final Answer:

   DECLARE cur_orders CURSOR FOR SELECT * FROM orders; OPEN cur_orders; LOOP FETCH cur_orders INTO rec; EXIT WHEN NOT FOUND; -- process rec END LOOP; CLOSE cur_orders; -> Option A

4. Quick Check:

   Declare, Open, Loop Fetch, Close = correct pattern [OK]

Hint: Use LOOP with FETCH and EXIT WHEN NOT FOUND [OK]

Common Mistakes:

• Opening cursor before declaring
• Fetching before opening
• Using FETCH ALL which is invalid for cursors