Migration from DataFaker

Use this mapping to move DataFaker (or JavaFaker) code to kRandom. This is feature-parity guidance, not source-compatible import replacement. JavaFaker users can use the same mappings: its original repository still documents the legacy 1.0.2 coordinate, while DataFaker is the actively developed successor.

Install

dependencies {
    implementation("io.github.frikit:krandom-core:2.10.0")
}

API mapping

DataFaker	kRandom equivalent
`new Faker()`	no instance needed — `Generators.of*()` statics or domain namespaces
`new Faker(Locale.GERMANY)`	`Generators.ofCity(Locale.GERMANY)` per call, or `GeneratorConfig.builder().locale(...)` shared
`new Faker(new Random(42))`	`GeneratorConfig.builder().seed(42L).build()`
`faker.name().fullName()`	`Generators.ofFullName().generate()`
`faker.name().firstName()`	`Generators.person().firstName().generate()`
`faker.name().lastName()`	`Generators.person().lastName().generate()`
`faker.internet().emailAddress()`	`Generators.ofEmail().generate()`
`faker.internet().url()`	`Generators.ofUrl().generate()`
`faker.internet().ipV4Address()`	`Generators.ofIPv4().generate()`
`faker.internet().username()`	`Generators.ofUsername().generate()`
`faker.address().city()`	`Generators.ofCity().generate()`
`faker.address().country()`	`Generators.ofCountry().generate()`
`faker.address().streetAddress()`	`Generators.ofStreetAddress().generate()`
`faker.address().zipCode()`	`Generators.ofPostalCode().generate()`
`faker.phoneNumber().phoneNumber()`	`Generators.ofPhoneNumber().generate()`
`faker.company().name()`	`Generators.ofCompanyName().generate()`
`faker.finance().iban()`	`new IbanGenerator(bankingConfig).generate()`
`faker.finance().creditCard()`	`Generators.ofCreditCard().generate()`
`faker.number().numberBetween(1, 100)`	`Generators.ofInt(1, 100).generate()`
`faker.regexify("[a-z]{8}")`	`new RegexGenerator("[a-z]{8}").generate()`
`faker.university().name()`	load a local data pack, then `Generators.ofUniversity(config).name()`

Domain namespaces (Generators.person(), location(), finance(), network(), text(), commerce(), identifier(), datetime()) mirror DataFaker’s provider-object style if you prefer fluent discovery over flat statics. Each accepts an optional GeneratorConfig.

In 2.0.0, banking identifiers are fail-closed. Define bankingConfig only for an isolated compatibility fixture; it does not make an IBAN safe to submit anywhere:

GeneratorConfig bankingConfig = GeneratorConfig.builder()
        .bankingSafetyPolicy(BankingSafetyPolicy.REALISTIC_UNCLASSIFIED)
        .build();

Template helpers

DataFaker’s string templates map directly:

DataFaker	kRandom equivalent
`faker.numerify("###-##")`	`Generators.ofTemplate("###-##").generate()`
`faker.letterify("????")`	`Generators.ofTemplate("????").generate()`
`faker.bothify("##??")`	`Generators.ofTemplate("##??").generate()`
`faker.expression("#{Name.firstName}")`	`Generators.ofDataFakerExpression("#{Name.firstName}").generate()` (same syntax) or `Generators.ofProviderTemplate("{firstname}").generate()` (native syntax)

ofProviderTemplate resolves {token} placeholders through the ProviderHub (tokens like {firstname}, {email}, {city}) and expands #/? placeholders in the same string — see the Schema and Provider Hub guide for the full token list.

Bulk and structured output

Both libraries have schema-based bulk output. kRandom’s v2 schema API streams generated rows to JSONL, JSON arrays, CSV, XML, SQL, YAML, and TOML, and exports JSON Schema. SchemaProjection<T> also transforms existing object sequences through those same formats without materializing a map batch. For generated row-style data use Field + Schema; for existing objects use SchemaProjection<T> — see Schema and Provider Hub.

Seeding differences

DataFaker shares one Random across the whole faker instance. kRandom seeds per generator or per shared GeneratorConfig; string seeds are supported via a stable derivation (GeneratorConfig.deriveSeed, fnv1a64-v1). Tests should assert on shape, not exact values — see VERSIONING.md for the seed-stability policy.

Krandom is the closest match for one seeded Faker instance: Krandom.seeded("name") returns plain values (fake.firstName(), fake.city(), fake.dateBetween(from, to)), and each method has its own stream, so a method’s values are unaffected by calls of other methods.

Honest gaps

Long-tail novelty providers (sports teams, movies, food, animals, pop-culture catalogs): not shipped. If your tests depend on these, keep DataFaker for those calls or open an issue — domain packs are community-contributable.
Locales: kRandom supports 50 variants (35 native datasets and 15 curated fallbacks), while DataFaker maintains a broader locale surface. Check the exact locale you use against the Locale-Aware Data guide.
Local data packs: DataFaker accepts YAML files and URLs. kRandom accepts a versioned, SHA-256-verified local University CSV pack with declared source and license; runtime network loading is intentionally excluded. See Local Data Packs.
#{...} expression syntax: the common tokens work directly via Generators.ofDataFakerExpression(...) (case-insensitive, camelCase or snake_case; unknown tokens fail fast with the supported list). Exotic provider tokens beyond that list still need manual mapping.