Quick Insight
Data integrity is the assurance that data remains accurate, complete, consistent, and trustworthy throughout its lifecycle, protected from accidental or unauthorized alteration. It covers physical integrity, which guards against hardware failure and corruption, and logical integrity, which is enforced through entity, referential, domain, and user-defined constraints in databases. It is closely related to, but distinct from, data quality and data security.
Types of Data Integrity
- Physical integrity: Data survives storage and transfer without corruption, power failures, or hardware faults.
- Entity integrity: Every record has a unique, non-empty identifier, so no two rows claim to be the same thing.
- Referential integrity: Links between records stay valid. A store record that points to a parent mall must point to a mall that actually exists.
- Domain integrity: Values stay within allowed ranges and formats, such as latitude between -90 and 90.
- User-defined integrity: Business rules, such as a closing date never coming before an opening date.
How Data Integrity Is Maintained
- Database constraints: Primary keys, foreign keys, and validation rules block bad writes.
- Checksums and hashes: Confirm a file hasn’t changed during transfer.
- Access controls and audit trails: Record who changed what and when.
- Validation at ingestion: Check incoming data before it enters production tables.
- Backups: Allow recovery from corruption or accidental deletion.
Regulated industries such as pharmaceuticals often follow ALCOA+ principles, which require records to be attributable, legible, contemporaneous, original, and accurate.
Data Integrity in Location Data
- Stable identifiers: A place should keep the same ID across monthly updates so history stays connected.
- Parent and child relationships: A store inside a shopping center needs a valid link to that center, part of the spatial hierarchy in POI data.
- Geometry matching its place: A building footprint must belong to the right business, or every visit or boundary analysis built on it is off.
How Data Integrity Compares
Term | Core question |
Data integrity | Has the data stayed accurate and unaltered? |
Data quality | Is the data fit for this specific use? |
Data security | Is the data protected from unauthorized access? |
A caveat: data can have perfect integrity and still be wrong if it was inaccurate when first recorded. Integrity protects data from change, not from bad input.
Frequently Asked Questions
1. What are the most common threats to data integrity?
Human error during entry or editing, failed transfers, software bugs, and malicious changes are the usual culprits.
2. How can I check whether a downloaded file is intact?
Compare its checksum or hash with the one published by the source. If they match, the file wasn’t altered or corrupted in transit.
3. Does encryption guarantee data integrity?
Not on its own. Encryption hides data, but integrity also needs checks like hashes or authenticated encryption to detect tampering.
4. Who is responsible for data integrity?
Everyone who touches the data, but data owners and stewards are typically accountable for setting and enforcing the rules.
5. Can spreadsheets maintain data integrity?
Only weakly. They lack enforced constraints and reliable audit trails, which is why critical data usually moves to databases.
