Schemas (Standard Schema)

Every identifier has a ready-made Standard Schema. Standard Schema is the common interface that Zod, Valibot, ArkType and others implement, and that form and API libraries accept directly. dr-utils implements it with zero dependencies.

Use schemas when a library asks you for "a schema". Use the plain validators everywhere else: they are unchanged and still the simplest option.

Schema	Accepts	Output (on success)
`cedulaSchema(options?)`	cédula, dashes and spaces allowed	`Cedula` — 11 digits
`rncSchema(options?)`	RNC or cédula, dashes and spaces allowed	`RNC` — 9 or 11 digits
`ncfSchema(options?)`	NCF / e-CF, surrounding whitespace ok	`NCF`
`phoneNumberSchema(options?)`	any common phone format, `+1` optional	`PhoneNumber` — 10 digits
`plateSchema(options?)`	license plate, any case	`Plate` — uppercase
`provinciaSchema(options?)`	province name, any case or accents	`Provincia` — official name
`municipioSchema(options?)`	municipality name, any case or accents	`Municipio` — official name

Options

Option	Type	Description
`message`	`string`	Error message for invalid input. Defaults to English (`'Invalid cedula'`).

Schemas validate synchronously and also reject non-string values.

With TanStack Form

TanStack Form accepts Standard Schemas as field validators:

import { useForm } from '@tanstack/react-form'
import { cedulaSchema, phoneNumberSchema } from 'dr-utils/schemas'

const form = useForm({ defaultValues: { cedula: '', telefono: '' } })

<form.Field
  name="cedula"
  validators={{ onChange: cedulaSchema({ message: 'Cédula inválida' }) }}
  children={(field) => (
    <>
      <input value={field.state.value} onChange={(e) => field.handleChange(e.target.value)} />
      <p>{field.state.meta.errors[0]?.message}</p>
    </>
  )}
/>

With any Standard Schema consumer

Any library that accepts a Standard Schema for a single value works the same way, for example a tRPC procedure input: .input(rncSchema()). You can also call a schema yourself:

import { rncSchema } from 'dr-utils/schemas'

const result = rncSchema()['~standard'].validate('130-72075-4')

if (result.issues) {
	console.log(result.issues[0].message) // 'Invalid RNC'
} else {
	console.log(result.value) // '130720754'
}

Inside Zod or Valibot objects

Zod and Valibot can't nest another library's schema inside their own objects. Use the validator functions with their refinement helpers instead; it's the same check:

import { z } from 'zod'
import { validateCedula, validatePhoneNumber } from 'dr-utils/validators'
import { normalizePhoneNumber } from 'dr-utils/parsers'

const Cliente = z.object({
	cedula: z.string().refine(validateCedula, 'Cédula inválida'),
	telefono: z
		.string()
		.refine(validatePhoneNumber, 'Teléfono inválido')
		.transform(normalizePhoneNumber), // '(809) 555-1234' → '8095551234'
})

import * as v from 'valibot'
import { validateRNC } from 'dr-utils/validators'

const Empresa = v.object({
	rnc: v.pipe(v.string(), v.check(validateRNC, 'RNC inválido')),
})

Types

import { cedulaSchema, type StandardSchemaV1 } from 'dr-utils'

type Output = StandardSchemaV1.InferOutput<ReturnType<typeof cedulaSchema>> // Cedula

With TanStack Form​

With any Standard Schema consumer​

Inside Zod or Valibot objects​

Types​

With TanStack Form

With any Standard Schema consumer

Inside Zod or Valibot objects

Types