the removal, replacement, or modification of unwanted or harmful data.
Simple user errors may result in helpful messaging indicating limited character sets,
minimum and maximum limits, and errors dictating additional user instruction. Attacks vary greatly in content and are treated quite differently from user errors.
Consider an SQL Database Query Injection, rather than an email address. Data Cleansing differs from Data Validation by removing, replacing, or modifying the submitted data to prevent any form of execution. Although the attack could be rejected, it may instead be rendered inert, appended to a log, and sent to an admin for review.
Data Cleansing renders attacks inert and defends against illegal data types.