UPDATE changes column values in existing rows; DELETE removes rows. Both touch every row the WHERE matches — which is all of them if you forget it.

UPDATE

SQL

-- 10% raise for Sales; several columns can change at once
UPDATE employees
SET salary = salary * 1.10,
    email  = COALESCE(email, lower(name) || '@corp.in')
WHERE dept_id = 2;

SELECT emp_id, name, email, salary FROM employees WHERE dept_id = 2 ORDER BY emp_id;

Output

emp_id | name  | email         | salary
-------+-------+---------------+-------------------
5      | Priya | priya@corp.in | 121000.00000000001
6      | Arjun | arjun@corp.in | 77000
7      | Neha  | neha@corp.in  | 79200

UPDATE using another table

SQL

-- raise everyone located in Hyderabad by 5000: subquery form (works everywhere)
UPDATE employees
SET salary = salary + 5000
WHERE dept_id IN (SELECT dept_id FROM departments WHERE location = 'Hyderabad');

SELECT name, salary FROM employees WHERE dept_id = 3 ORDER BY emp_id;

Output

name   | salary
-------+-------
Vikram | 85000
Sara   | 90000

SQL

-- UPDATE … FROM (PostgreSQL, SQLite 3.33+): join the table being updated
CREATE TABLE raises (emp_id INTEGER PRIMARY KEY, pct INTEGER);
INSERT INTO raises VALUES (3, 10), (4, 20);

UPDATE employees
SET salary = salary + salary * r.pct / 100
FROM raises AS r
WHERE r.emp_id = employees.emp_id;

SELECT name, salary FROM employees WHERE emp_id IN (3, 4) ORDER BY emp_id;

Output

name  | salary
------+-------
Meera | 104500
Karan | 114000

MySQL / SQL Server

-- MySQL: multi-table UPDATE with JOIN
UPDATE employees e
JOIN raises r ON r.emp_id = e.emp_id
SET e.salary = e.salary + e.salary * r.pct / 100;

-- SQL Server: UPDATE … FROM with an alias
UPDATE e SET e.salary = e.salary + e.salary * r.pct / 100
FROM employees e JOIN raises r ON r.emp_id = e.emp_id;

DELETE

SQL

-- remove order lines of cancelled orders, then the orders themselves
DELETE FROM order_items
WHERE order_id IN (SELECT order_id FROM orders WHERE status = 'cancelled');

DELETE FROM orders WHERE status = 'cancelled'
RETURNING order_id, customer_id;

SELECT COUNT(*) AS orders_left FROM orders;

Output

order_id | customer_id
---------+------------
105      | 4

orders_left
-----------
7

The child rows (order_items) had to go first — otherwise the foreign key would block deleting the order (unless it was declared ON DELETE CASCADE).

Safe habits

Preview
Run SELECT … WHERE <same condition> first and check the row count.
Wrap in a transaction
BEGIN; UPDATE …; → check the result → COMMIT or ROLLBACK.
Limit the blast radius
Key the WHERE on the primary key when you mean one row.
Prefer soft deletes for business data
UPDATE … SET deleted_at = CURRENT_TIMESTAMP keeps history and is reversible.

SQL

BEGIN;
UPDATE employees SET salary = 1;          -- oops: no WHERE, all 10 rows changed
SELECT COUNT(*) AS paid_one_rupee FROM employees WHERE salary = 1;
ROLLBACK;                                 -- undo everything since BEGIN
SELECT name, salary FROM employees WHERE emp_id = 1;

Output

paid_one_rupee
--------------
10

name | salary
-----+-------
Asha | 150000

How do you delete rows from one table based on another table?

What happens if you run UPDATE without WHERE?