A hash value is the fixed-size output produced when a hash function transforms data. Deterministic hashing gives repeated inputs the same value, allowing a hash to serve as a pseudonymous identifier for joining records without exposing the original field. section 8.3.2
Use hash values with these boundaries in mind:
- Treat the value as an identifier, not as a measure of similarity: small changes in the original data produce very different hashes. section 8.3.2
- Keep the original sensitive data in a separate, restricted table while exposing only the hash in the analytics table. section 8.3.2
- Do not assume reversal is permanently impossible; review current information-security practices before relying on hashing for protection. section 8.3.2
Hashing email addresses in pandas
Use Python’s hashlib library and SHA-256 to create an email_hash column, then save the original email and hash separately and remove email from the transactions table. section 8.3.2
