jt.dump and jt.dumpJson

dump and dumpJson are a symmetric pair with instantiate: instantiate ingests and decodes wire data; dump / dumpJson encodes domain data back to wire form.

---

jt.dump

Declaration. Walks the canonical schema graph for the given schemaId, applies any registered Transform encoder at each node, filters the result according to options, and returns a wire-form JS value (unknown). The input value is not mutated.

Options

Option	Type	Default	Description
mode	'wire' | 'json'	'wire'	'json' converts Date values to ISO strings for JSON.stringify safety
exclude	readonly string[]	-	Property names to drop (ignored when include is set)
include	readonly string[]	-	Property names to keep (takes precedence over exclude)
excludeUnset	boolean	false	Drop properties whose runtime value is undefined
excludeDefaults	boolean	false	Drop properties whose value strictly equals the schema default

Use this when you need to serialize a domain object back to wire form - before storing in a database, before sending over HTTP, before publishing to a queue. Use the filtering options to produce compact payloads or specific projections. Use dumpJson when you need a JSON string directly.

Don't use this when you want a complete validated object (use instantiate instead - it goes the other direction). Don't call it on raw unvalidated input - dump expects a value that has already been through instantiate or materialize.

Examples

Example 1: Basic serialization of a coerced book

import { bookstoreEntities as entities, BookSchema } from './bookstore/index.js';

const book = jt.instantiate(BookSchema.$id, {
  isbn:    '9780140449136',
  title:   'Crime and Punishment',
  authors: ['Fyodor Dostoevsky'],
  price:   14.99,
  // currency defaults to 'USD', inStock defaults to true
});

// Wire-form output  - all fields including defaults
const wire = jt.dump(BookSchema.$id, book);
// { isbn: '...', title: '...', authors: [...], price: 14.99, currency: 'USD', inStock: true }

Example 2: Compact payload - exclude default-valued fields

const compact = jt.dump(BookSchema.$id, book, { excludeDefaults: true });
// { isbn: '...', title: '...', authors: [...], price: 14.99 }
// currency: 'USD' and inStock: true are omitted (equal to schema defaults)

Example 3: Project to specific fields

const listing = jt.dump(BookSchema.$id, book, { include: ['isbn', 'title', 'price'] });
// { isbn: '9780140449136', title: 'Crime and Punishment', price: 14.99 }

Example 4: Transform integration - encode applied automatically

If the schema has a Transform encoder registered (see Transforms), dump applies the encode function at each transformed node. A instantiate → dump round-trip recovers the original wire value.

import { Transform } from 'json-tology';

const PlacedAtSchema = Transform.create(
  { $id: 'https://bookstore.example/PlacedAt', type: 'string', format: 'date-time' } as const,
  { decode: (s: string) => new Date(s), encode: (d: Date) => d.toISOString() },
);

const date = jt.instantiate(PlacedAtSchema.$id, '2026-01-15T10:30:00.000Z');
const wire = jt.dump(PlacedAtSchema.$id, date as Date);
// '2026-01-15T10:30:00.000Z'  - encode applied automatically

Bad examples - what NOT to do

Anti-pattern 1: Calling dump on raw (uninstantiated) input

// ⊥ Don't do this  - dump expects an instantiated domain value, not raw input
const raw = { isbn: '9780140449136', title: 'Crime...', authors: ['Dostoevsky'], price: '14.99' };
const wire = jt.dump(BookSchema.$id, raw as Book);
// price is still '14.99' string  - not coerced; dump just applies encode, not type coercion

// ✓ Do this  - instantiate first, then dump
const book = jt.instantiate(BookSchema.$id, raw);
const wireBook = jt.dump(BookSchema.$id, book);

Comparison

json-tology

const wire = jt.dump(BookSchema.$id, book);
const wire2 = jt.dump(BookSchema.$id, book, { excludeDefaults: true });
const json = jt.dumpJson(BookSchema.$id, book); // JSON string

Zod

// Zod doesn't have a separate dump/serialize step.
// The parsed value is already in the desired shape.
const json = JSON.stringify(book);
// No exclude/include filtering without manual code.

Valibot

// Limitation: Valibot has no dump/serialize step and no encode direction.
// JSON.stringify the value directly; encoding of branded or transformed
// fields (e.g. Date) must be applied manually before stringify.
const json = JSON.stringify(book);

TypeBox + Value

// TypeBox does not have a built-in dump/serialize utility.
const json = JSON.stringify(book);

AJV

// AJV validates  - no serialize step.
const json = JSON.stringify(book);

Pydantic

wire = book.model_dump()
json_str = book.model_dump_json()
# Filtering options:
book.model_dump(exclude={'currency', 'in_stock'})
book.model_dump(include={'isbn', 'title', 'price'})
book.model_dump(exclude_defaults=True)
book.model_dump(exclude_none=True)

---

jt.dumpJson

Declaration. Convenience wrapper around dump() with mode: 'json' forced. Equivalent to JSON.stringify(jt.dump(schemaId, value, { mode: 'json', ...options })). Returns a JSON string. The mode option is not available on dumpJson - it is always 'json'.

Use this when you need a JSON string directly - HTTP response bodies, log records, message queue payloads.

Examples

Example 1: Serialize a customer for an HTTP response

const customer = jt.instantiate(CustomerSchema.$id, {
  id:    'c1a2b3d4-e5f6-7890-abcd-ef1234567890',
  email: 'alice@bookstore.example',
  name:  'Alice Chen',
});

const json = jt.dumpJson(CustomerSchema.$id, customer);
// '{"id":"...","email":"alice@bookstore.example","name":"Alice Chen","addresses":[]}'

// In an Express handler:
// res.type('application/json').send(json);

Example 2: Compact order payload

const orderJson = jt.dumpJson(OrderSchema.$id, order, { excludeDefaults: true });
// JSON string without currency: 'USD'

Comparison

json-tology

const json = jt.dumpJson(CustomerSchema.$id, customer);
// JSON string; Date fields encoded to ISO via Transform encoders

Zod

const json = JSON.stringify(customer);
// Date objects must be converted manually before stringify.

Valibot

// Limitation: no dumpJson; JSON.stringify works only for plain values.
const json = JSON.stringify(customer);

TypeBox + Value

const json = JSON.stringify(customer);

AJV

const json = JSON.stringify(customer);

Pydantic

json_str = customer.model_dump_json()
# Equivalent; datetime fields serialized to ISO automatically.

Related

• JsonTology.instantiate - the incoming direction (wire → domain)
• jt.encode - apply a single Transform encoder
• Transforms - how Transform encoders are registered and applied

See also

• Bookstore domain - where Book, Customer, Order are defined