Data Validation vs Data Cleaning

"Clean the data" and "validate the data" get used interchangeably, but they're different operations with different jobs. Confusing them leads to pipelines that either mangle records they should have rejected or reject records they should have fixed.

The distinction

Validation: enforce a contract at the boundary

Use Guard when you need to enforce rules on inbound data — a webhook payload, an upload, a form. It returns a pass/fail per record with the specific reason a field failed, so bad data stops at the door instead of poisoning a table.

curl -X POST https://zapinner.com/api/v1/guard \
  -H "Authorization: Bearer $ZAPINNER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "records": [{ "email": "not-an-email" }], "expect": { "email": "email" } }'
# -> { passed: false, failures: [{ field: "email", reason: "invalid_format" }] }

Cleaning: fix the records you accept

Use Repair or Normalize when the right move is to fix a value rather than reject the record — casing, formatting, parsing. Cleaning changes data and returns an audit of what it changed.

Use them together

The strongest pipelines do both: Map incoming data into your schema, Guard it against your contract, and Repair the records you choose to accept. Validate to decide, clean to fix.