Data Normalization Explained

Normalization is the unglamorous step that makes everything downstream possible: turning the same value expressed a dozen ways into one canonical form. It sounds trivial until you're the one maintaining the parsing code for every source.

What it means field by field

Why it has to be deterministic

If the same input can normalize two different ways, you can't cache results, can't test them, and can't trust that a record processed today matches one processed last week. Zapinner's normalize operation is deterministic by design — the same input always produces the same output.

Type-directed normalization

You tell Normalize the intended type of each field and it canonicalizes accordingly, rather than guessing. That makes the result predictable and the request self-documenting.

curl -X POST https://zapinner.com/api/v1/normalize \
  -H "Authorization: Bearer $ZAPINNER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "records": [{ "email": " JANE@ACME.COM ", "revenue": "$1.2M" }],
    "types": { "email": "email", "revenue": "currency" }
  }'

Where it fits

Normalize early — right as data enters your system — so every later step (matching, validation, storage) works on one consistent shape. It's also the first engine Repair runs internally, which is why Repair's output is consistent regardless of how messy the input was.