This is a model-driven approach to handling Forms inputs and validations, heavily inspired in Angular's Reactive Forms.
- Getting Started
- Creating a form
- How to get/set Form data
- Validators
- Groups of Groups
- Dynamic forms with FormArray
- Arrays of Groups
- FormBuilder
- Reactive Form Widgets
- How to customize error messages?
- When does Validation Messages begin to show up?
- Enable/Disable Submit button
- Focus/UnFocus a FormControl
- Focus flow between Text Fields
- Enable/Disable a widget
- How does ReactiveTextField differs from native TextFormField or TextField?
- Reactive Form Field Widgets
- Bonus Field Widgets
- Other Reactive Forms Widgets
- Advanced Reactive Field Widgets
- ReactiveValueListenableBuilder to listen when value changes in a FormControl
- ReactiveForm vs ReactiveFormBuilder which one?
- Reactive Forms + Provider plugin
- Reactive Forms + code generation plugin
- How create a custom Reactive Widget?
- What is not Reactive Forms
- What is Reactive Forms
- Migrate versions
For help getting started with Flutter, view the online documentation, which offers tutorials, samples, guidance on mobile development, and a full API reference.
- Dart SDK: >=3.2.0 <4.0.0
- Flutter: >=3.16.0
For using Reactive Forms in projects below Flutter 2.8.0 please use the version <= 10.7.0 of Reactive Forms.
For using Reactive Forms in projects below Flutter 2.2.0 please use the version <= 10.2.0 of Reactive Forms.
For using Reactive Forms in projects with Flutter 1.17.0 please use the version 7.6.3 of Reactive Forms.
Reactive Forms v8.x includes the intl package. If a version conflict is present, then you should use dependency_overrides to temporarily override all references to intl and set the one that better fits your needs.
Once you're familiar with Flutter you may install this package adding reactive_forms to the dependencies list
of the pubspec.yaml file as follow:
dependencies:
flutter:
sdk: flutter
reactive_forms: ^17.0.1Then run the command flutter packages get on the console.
A form is composed by multiple fields or controls.
To declare a form with the fields name and email is as simple as:
final form = FormGroup({
'name': FormControl<String>(value: 'John Doe'),
'email': FormControl<String>(),
});Notice in the example above that in the case of the name we have also set a default value, in the case of the email the default value is null.
Given the FormGroup:
final form = FormGroup({
'name': FormControl<String>(value: 'John Doe'),
'email': FormControl<String>(value: '[email protected]'),
});You can get the value of a single FormControl as simple as:
String get name() => this.form.control('name').value;But you can also get the complete Form data as follows:
print(form.value);The previous code prints the following output:
{
"name": "John Doe",
"email": "[email protected]"
}FormGroup.value returns an instance of Map<String, dynamic> with each field and its value.
To set value to controls you can use two approaches:
// set value directly to the control
this.form.control('name').value = 'John';
// set value to controls by setting value to the form
this.form.value = {
'name': 'John',
'email': '[email protected]',
};You can add validators to a FormControl as follows:
final form = FormGroup({
'name': FormControl<String>(validators: [Validators.required]),
'email': FormControl<String>(validators: [
Validators.required,
Validators.email,
]),
});If at least one FormControl is invalid then the FormGroup is invalid
There are common predefined validators, but you can implement custom validators too.
- Validators.required
- Validators.requiredTrue
- Validators.email
- Validators.number
- Validators.min
- Validators.max
- Validators.minLength
- Validators.maxLength
- Validators.pattern
- Validators.creditCard
- Validators.equals
- Validators.compose
- Validators.composeOR
- Validators.any
- Validators.contains
- Validators.mustMatch
- Validators.compare
- Validators.minLength
- Validators.maxLength
- Validators.any
- Validators.contains
All validators are instances of classes that inherit from the Validator abstract class.
In order to implement a custom validator you can follow two different approaches:
1- Extend from Validator class and override the validate method.
2- Or implement a custom validator function|method, and use it with the Validators.delegate(...) validator.
Let's implement a custom validator that validates a control's value must be true:
Let's create a class that extends from Validator and overrides the validate method:
/// Validator that validates the control's value must be `true`.
class RequiredTrueValidator extends Validator<dynamic> {
const RequiredTrueValidator() : super();
@override
Map<String, dynamic>? validate(AbstractControl<dynamic> control) {
return control.isNotNull &&
control.value is bool &&
control.value == true
? null
: {'requiredTrue': true};
}
}The validator method is a function that receives the control to validate and returns a Map. If the value of the control is valid the function returns null, otherwise returns a Map with the error key and a custom information. In the previous example we have defined requiredTrue as the error key and true as the custom information.
In order to use the new validator class we provide an instance of it in the FormControl definition.
final form = FormGroup({
'acceptLicense': FormControl<bool>(
value: false,
validators: [
RequiredTrueValidator(), // providing the new custom validator
],
),
});Sometimes it's more convenient to implement a custom validator in a separate method|function than in a different new class. In that case, it is necessary to use the Validators.delegate() validator. It creates a validator that delegates the validation to the external function|method.
final form = FormGroup({
'acceptLicense': FormControl<bool>(
value: false,
validators: [
Validators.delegate(_requiredTrue) // delegates validation to a custom function
],
),
});/// Custom function that validates that control's value must be `true`.
Map<String, dynamic>? _requiredTrue(AbstractControl<dynamic> control) {
return control.isNotNull &&
control.value is bool &&
control.value == true
? null
: {'requiredTrue': true};
}Check the Migration Guide to learn more about custom validators after version 15.0.0 of the package.
Validator.pattern is a validator that comes with Reactive Forms. Validation using regular expressions have been always a very useful tool to solve validation requirements. Let's see how we can validate American Express card numbers:
American Express card numbers start with 34 or 37 and have 15 digits.
const americanExpressCardPattern = r'^3[47][0-9]{13}$';
final cardNumber = FormControl<String>(
validators: [Validators.pattern(americanExpressCardPattern)],
);
cardNumber.value = '395465465421'; // not a valid number
expect(cardNumber.valid, false);
expect(cardNumber.hasError('pattern'), true);The above code is a Unit Test extracted from Reactive Forms tests.
If we print the value of FormControl.errors:
print(cardNumber.errors);We will get a Map like this:
{
"pattern": {
"requiredPattern": "^3[47][0-9]{13}$",
"actualValue": 395465465421
}
}There are special validators that can be attached to FormGroup. In the next section we will see an example of that.
There are some cases where we want to implement a Form where a validation of a field depends on the value of another field. For example a sign-up form with email and emailConfirmation or password and passwordConfirmation.
For those cases we could implement a custom validator as a class and attach it to the FormGroup. Let's see an example:
final form = FormGroup({
'name': FormControl<String>(validators: [Validators.required]),
'email': FormControl<String>(validators: [Validators.required, Validators.email]),
'password': FormControl<String>(validators: [
Validators.required,
Validators.minLength(8),
]),
'passwordConfirmation': FormControl<String>(),
}, validators: [
MustMatchValidator(controlName: 'password', matchingControlName: 'passwordConfirmation')
]);Notice the use of *Validators.minLength(8)*
In the previous code we have added two more fields to the form: password and passwordConfirmation, both fields are required and the password must be at least 8 characters length.
However the most important thing here is that we have attached a validator to the FormGroup. This validator is a custom validator and the implementation follows as:
class MustMatchValidator extends Validator<dynamic> {
final String controlName;
final String matchingControlName;
MustMatchValidator({
required this.controlName,
required this.matchingControlName,
}) : super();
@override
Map<String, dynamic>? validate(AbstractControl<dynamic> control) {
final form = control as FormGroup;
final formControl = form.control(controlName);
final matchingFormControl = form.control(matchingControlName);
if (formControl.value != matchingFormControl.value) {
matchingFormControl.setErrors({'mustMatch': true});
// force messages to show up as soon as possible
matchingFormControl.markAsTouched();
} else {
matchingFormControl.removeError('mustMatch');
}
return null;
}
}Fortunately you don't have to implement a custom must match validator because we have already included it into the code of the reactive_forms package so you should reuse it. The previous form definition becomes into:
final form = FormGroup({
'name': FormControl<String>(validators: [Validators.required]),
'email': FormControl<String>(validators: [Validators.required, Validators.email]),
'emailConfirmation': FormControl<String>(),
'password': FormControl<String>(validators: [Validators.required, Validators.minLength(8)]),
'passwordConfirmation': FormControl<String>(),
}, validators: [
Validators.mustMatch('email', 'emailConfirmation'),
Validators.mustMatch('password', 'passwordConfirmation'),
]);Some times you want to perform a validation against a remote server, this operations are more time consuming and need to be done asynchronously.
For example you want to validate that the email the user is currently typing in a registration form is unique and is not already used in your application. Asynchronous Validators are just another tool so use them wisely.
Asynchronous Validators are very similar to their synchronous counterparts, with the following difference:
- The validator function returns a Future
Asynchronous validation executes after the synchronous validation, and is performed only if the synchronous validation is successful. This check allows forms to avoid potentially expensive async validation processes (such as an HTTP request) if the more basic validation methods have already found invalid input.
After asynchronous validation begins, the form control enters a pending state. You can inspect the control's pending property and use it to give visual feedback about the ongoing validation operation.
Code speaks more than a thousand words :) so let's see an example.
Let's implement the previous mentioned example: the user is typing the email in a registration Form and you want to validate that the email is unique in your System. We will implement a custom async validator for that purpose.
final form = FormGroup({
'email': FormControl<String>(
validators: [
Validators.required, // traditional required and email validators
Validators.email,
],
asyncValidators: [
UniqueEmailAsyncValidator(), // custom asynchronous validator :)
],
),
});We have declared a simple Form with an email field that is required and must have a valid email value, and we have include a custom async validator that will validate if the email is unique. Let's see the implementation of our new async validator:
/// Validator that validates the user's email is unique, sending a request to
/// the Server.
class UniqueEmailAsyncValidator extends AsyncValidator<dynamic> {
@override
Future<Map<String, dynamic>?> validate(AbstractControl<dynamic> control) async {
final error = {'unique': false};
final isUniqueEmail = await _getIsUniqueEmail(control.value.toString());
if (!isUniqueEmail) {
control.markAsTouched();
return error;
}
return null;
}
/// Simulates a time consuming operation (i.e. a Server request)
Future<bool> _getIsUniqueEmail(String email) {
// simple array that simulates emails stored in the Server DB.
final storedEmails = ['[email protected]', '[email protected]'];
return Future.delayed(
const Duration(seconds: 5),
() => !storedEmails.contains(email),
);
}
}Note the use of control.markAsTouched() to force the validation message to show up as soon as possible.
The previous implementation was a simple validator that receives the AbstractControl and returns a Future that completes 5 seconds after its call and performs a simple check: if the value of the control is contained in the server array of emails.
If you want to see Async Validators in action with a full example using widgets and animations to feedback the user we strong advice you to visit our Wiki. We have not included the full example in this README.md file just to simplify things here and to not anticipate things that we will see later in this doc.
The validator
Validators.delegateAsync()is another way to implement custom validator, for more reference check the Custom validators section.
Asynchronous validators have a debounce time that is useful if you want to minimize requests to a remote API. The debounce time is set in milliseconds and the default value is 250 milliseconds.
You can set a different debounce time as an optionally argument in the FormControl constructor.
final control = FormControl<String>(
asyncValidators: [UniqueEmailAsyncValidator()],
asyncValidatorsDebounceTime: 1000, // sets 1 second of debounce time.
);To explain what Composing Validators is, let's see an example:
We want to validate a text field of an authentication form. In this text field the user can write an email or a phone number and we want to make sure that the information is correctly formatted. We must validate that input is a valid email or a valid phone number.
final phonePattern = '<some phone regex pattern>';
final form = FormGroup({
'user': FormControl<String>(
validators: [
Validators.composeOR([
Validators.email,
Validators.pattern(phonePattern),
])
],
),
});Note that Validators.composeOR receives a collection of validators as argument and returns a validator.
With Validators.composeOR we are saying to FormControl that if at least one validator evaluate as VALID then the control is VALID it's not necessary that both validators evaluate to valid.
Another example could be to validate multiples Credit Card numbers. In that case you have several regular expression patterns for each type of credit card. So the user can introduce a card number and if the information match with at least one pattern then the information is considered as valid.
final form = FormGroup({
'cardNumber': FormControl<String>(
validators: [
Validators.composeOR([
Validators.pattern(americanExpressCardPattern),
Validators.pattern(masterCardPattern),
Validators.pattern(visaCardPattern),
])
],
),
});FormGroup is not restricted to contains only FormControl, it can nest others FormGroup so you can create more complex Forms.
Supose you have a Registration Wizzard with several screens. Each screen collect specific information and at the end you want to collect all that information as one piece of data:
final form = FormGroup({
'personal': FormGroup({
'name': FormControl<String>(validators: [Validators.required]),
'email': FormControl<String>(validators: [Validators.required]),
}),
'phone': FormGroup({
'phoneNumber': FormControl<String>(validators: [Validators.required]),
'countryIso': FormControl<String>(validators: [Validators.required]),
}),
'address': FormGroup({
'street': FormControl<String>(validators: [Validators.required]),
'city': FormControl<String>(validators: [Validators.required]),
'zip': FormControl<String>(validators: [Validators.required]),
}),
});Note how we have set the data type to a FormControl, although this is not mandatory when declaring a Form, we highly recommend this syntax as good practice or to use the FormBuilder syntax.
Using FormBuilder (read FormBuilder section below):
final form = fb.group({
'personal': fb.group({
'name': ['', Validators.required],
'email': ['', Validators.required],
}),
'phone': fb.group({
'phoneNumber': ['', Validators.required],
'countryIso': ['', Validators.required],
}),
'address': fb.group({
'street': ['', Validators.required],
'city': ['', Validators.required],
'zip': ['', Validators.required],
}),
});You can collect all data using FormGroup.value:
void _printFormData(FormGroup form) {
print(form.value);
}The previous method outputs a Map as the following one:
{
"personal": {
"name": "...",
"email": "..."
},
"phone": {
"phoneNumber": "...",
"countryIso": "..."
},
"address": {
"street": "...",
"city": "...",
"zip": "..."
}
}And of course you can access to a nested FormGroup as following:
FormGroup personalForm = form.control('personal');A simple way to create a wizard is for example to wrap a PageView within a ReactiveForm and each Page inside the PageView can contains a ReactiveForm to collect specific data.
FormArray is an alternative to FormGroup for managing any number of unnamed controls. As with FormGroup instances, you can dynamically insert and remove controls from FormArray instances, and the form array instance value and validation status is calculated from its child controls.
You don't need to define a key for each control by name, so this is a great option if you don't know the number of child values in advance.
Let's see a simple example:
final form = FormGroup({
'emails': FormArray<String>([]), // an empty array of emails
});We have defined just an empty array. Let's define another array with two controls:
final form = FormGroup({
'emails': FormArray<String>([
FormControl<String>(value: '[email protected]'),
FormControl<String>(value: '[email protected]'),
]),
});Note that you don't have to specify the name of the controls inside of the array.
If we output the value of the previous form group we will get something like this:
print(form.value);{
"emails": ["[email protected]", "[email protected]"]
}Let's dynamically add another control:
final array = form.control('emails') as FormArray<String>;
// adding another email
array.add(
FormControl<String>(value: '[email protected]'),
);
print(form.value);{
"emails": ["[email protected]", "[email protected]", "[email protected]"]
}Another way of add controls is to assign values directly to the array:
// Given: an empty array of strings
final array = FormArray<String>([]);
// When: set value to array
array.value = ["[email protected]", "[email protected]", "[email protected]"];
// Then: the array is no longer empty
expect(array.controls.length, 3);
// And: array has a control for each inserted value
expect(array.controls('0').value, "[email protected]");
expect(array.controls('1').value, "[email protected]");
expect(array.controls('2').value, "[email protected]");To get a control from the array you must pass the index position as a String. This is because FormGroup and FormArray inherited from the same parent class and FormControl gets the controls by name (String).
A more advanced example:
// an array of contacts
final contacts = ['[email protected]', '[email protected]', '[email protected]'];
// a form with a list of selected emails
final form = FormGroup({
'selectedEmails': FormArray<bool>([], // an empty array of controls
validators: [emptyAddressee], // validates that at least one email is selected
),
});
// get the array of controls
final formArray = form.control('selectedEmails') as FormArray<bool>;
// populates the array of controls.
// for each contact add a boolean form control to the array.
formArray.addAll(
contacts.map((email) => FormControl<bool>(value: true)).toList(),
);// validates that at least one email is selected
Map<String, dynamic> emptyAddressee(AbstractControl control) {
final emails = (control as FormArray<bool>).value;
return emails.any((isSelected) => isSelected)
? null
: {'emptyAddressee': true};
}You can also create arrays of groups:
// an array of groups
final addressArray = FormArray([
FormGroup({
'city': FormControl<String>(value: 'Sofia'),
'zipCode': FormControl<int>(value: 1000),
}),
FormGroup({
'city': FormControl<String>(value: 'Havana'),
'zipCode': FormControl<int>(value: 10400),
}),
]);Another example using FormBuilder:
// an array of groups using FormBuilder
final addressArray = fb.array([
fb.group({'city': 'Sofia', 'zipCode': 1000}),
fb.group({'city': 'Havana', 'zipCode': 10400}),
]);or just:
// an array of groups using a very simple syntax
final addressArray = fb.array([
{'city': 'Sofia', 'zipCode': 1000},
{'city': 'Havana', 'zipCode': 10400},
]);You can iterate over groups as follow:
final cities = addressArray.controls
.map((control) => control as FormGroup)
.map((form) => form.control('city').value)
.toList();A common mistake is to declare an array of groups as FormArray<FormGroup>.
An array of FormGroup must be declared as FormArray() or as FormArray<Map<String, dynamic>>().
The FormBuilder provides syntactic sugar that shortens creating instances of a FormGroup, FormArray and FormControl. It reduces the amount of boilerplate needed to build complex forms.
// creates a group
final form = fb.group({
'name': 'John Doe',
'email': ['', Validators.required, Validators.email],
'password': Validators.required,
});The previous code is equivalent to the following one:
final form = FormGroup({
'name': FormControl<String>(value: 'John Doe'),
'email': FormControl<String>(value: '', validators: [Validators.required, Validators.email]),
'password': FormControl<String>(validators: [Validators.required]),
});// creates an array
final aliases = fb.array(['john', 'little john']);// creates a control of type String with a required validator
final control = fb.control<String>('', [Validators.required]);// create a group
final group = fb.group(
// creates a control with default value and disabled state
'name': fb.state(value: 'john', disabled: true),
);To retrieves nested controls you can specify the name of the control as a dot-delimited string that define the path to the control:
final form = FormGroup({
'address': FormGroup({
'city': FormControl<String>(value: 'Sofia'),
'zipCode': FormControl<int>(value: 1000),
}),
});
// get nested control value
final city = form.control('address.city');
print(city.value); // outputs: SofiaSo far we have only defined our model-driven form, but how do we bind the form definition with our Flutter widgets? Reactive Forms Widgets is the answer ;)
Let's see an example:
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: Column(
children: <Widget>[
ReactiveTextField(
formControlName: 'name',
),
ReactiveTextField(
formControlName: 'email',
),
ReactiveTextField(
formControlName: 'password',
obscureText: true,
),
],
),
);
}The example above ignores the emailConfirmation and passwordConfirmation fields previously seen for simplicity.
Validation messages can be defined at two different levels:
- Reactive Widget level.
- Global/Application level.
Each reactive widget like ReactiveTextField, ReactiveDropdownField, and all others have the
property validationMessages as an argument of their constructors. In order to define custom
validation messages at widget level, just provide the property validationMessages with the
corresponding text values for each error as shown below:
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: Column(
children: <Widget>[
ReactiveTextField(
formControlName: 'name',
validationMessages: {
'required': (error) => 'The name must not be empty'
},
),
ReactiveTextField(
formControlName: 'email',
validationMessages: {
'required': (error) => 'The email must not be empty',
'email': (error) => 'The email value must be a valid email'
},
),
ReactiveTextField(
formControlName: 'password',
obscureText: true,
validationMessages: {
'required': (error) => 'The password must not be empty',
'minLength': (error) => 'The password must have at least 8 characters'
},
),
],
),
);
}Reactive Forms have an utility class called ValidationMessage that brings access to common validation messages: required, email, pattern and so on. So instead of write 'required' you could use ValidationMessage.required as the key of validation messages:
return ReactiveTextField( formControlName: 'email', validationMessages: { ValidationMessage.required: (error) => 'The email must not be empty', ValidationMessage.email: (error) => 'The email value must be a valid email', }, ),nice isn't it? ;)
You can also define custom validation messages at a higher level, for example, at the application level. When a reactive widget looks for an error message text, it first looks at widget level definition, if it doesn't find any config at widget level then it looks at the global config definition.
The global definition of validation messages allows you to define error messages in a centralized way and relieves you to define validation messages on each reactive widget of your application.
In order to define these configs at a higher level use the widget ReactiveFormConfig and
define the validationMessages.
Here is an example of the global definition for custom validation messages:
class MyApp extends StatelessWidget {
const MyApp({super.key});
@override
Widget build(BuildContext context) {
return ReactiveFormConfig(
validationMessages: {
ValidationMessage.required: (error) => 'Field must not be empty',
ValidationMessage.email: (error) => 'Must enter a valid email',
},
child: MaterialApp(
home: Scaffold(
body: const Center(
child: Text('Hello Flutter Reactive Forms!'),
),
),
),
);
}
}You can enrich the validation messages using parameters of the error instance. In the next example we are giving a more complete validation error to the user:
final form = FormGroup({
'password': FormControl<String>(
validators: [Validators.minLength(8)],
),
});ReactiveTextField(
formControlName: 'password',
validationMessage: {
ValidationMessages.minLength: (error) =>
'The password must be at least ${(error as Map)['requiredLength']} characters long'
},
)This will show the message: The password must be at least 8 characters long
Even when the FormControl is invalid, validation messages will begin to show up when the FormControl is touched. That means when the user taps on the ReactiveTextField widget and then remove focus or completes the text edition.
You can initialize a FormControl as touched to force the validation messages to show up at the very first time the widget builds.
final form = FormGroup({
'name': FormControl<String>(
value: 'John Doe',
validators: [Validators.required],
touched: true,
),
});When you set a value to a FormControl from code and want to show up validations messages you must call FormControl.markAsTouched() method:
set name(String newName) {
final formControl = this.form.control('name');
formControl.value = newName;
formControl.markAsTouched();// if newName is invalid then validation messages will show up in UI
}To mark all children controls of a FormGroup and FormArray you must call markAllAsTouched().
final form = FormGroup({ 'name': FormControl<String>( value: 'John Doe', validators: [Validators.required], touched: true, ), }); // marks all children as touched form.markAllAsTouched();
The second way to customize when to show error messages is to override the method showErrors in reactive widgets.
Let's suppose you want to show validation messages not only when it is invalid and touched (default behavior), but also when it's dirty:
ReactiveTextField(
formControlName: 'email',
// override default behavior and show errors when: INVALID, TOUCHED and DIRTY
showErrors: (control) => control.invalid && control.touched && control.dirty,
),A control becomes dirty when its value change through the UI.
The method setErrors of the controls can optionally mark it as dirty too.
For a better User Experience some times we want to enable/disable the Submit button based on the validity of the Form. Getting this behavior, even in such a great framework as Flutter, some times can be hard and can lead to have individual implementations for each Form of the same application plus boilerplate code.
We will show you two different approaches to accomplish this very easily:
- Separating Submit Button in a different Widget.
- Using ReactiveFormConsumer widget.
Let's add a submit button to our Form:
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: Column(
children: <Widget>[
ReactiveTextField(
formControlName: 'email',
),
ReactiveTextField(
formControlName: 'password',
obscureText: true,
),
MySubmitButton(),
],
),
);
}The above is a simple sign-in form with email, password, and a submit button.
Now let's see the implementation of the MySubmitButton widget:
class MySubmitButton extends StatelessWidget {
@override
Widget build(BuildContext context) {
final form = ReactiveForm.of(context);
return RaisedButton(
child: Text('Submit'),
onPressed: form.valid ? _onPressed : null,
);
}
void _onPressed() {
print('Hello Reactive Forms!!!');
}
}Notice the use of ReactiveForm.of(context) to get access to the nearest FormGroup up the widget's tree.
In the previous example we have separated the implementation of the submit button in a different widget. The reasons behind this is that we want to re-build the submit button each time the validity of the FormGroup changes. We don't want to rebuild the entire Form, but just the button.
How is that possible? Well, the answer is in the expression:
final form = ReactiveForm.of(context);The expression above have two important responsibilities:
- Obtains the nearest FormGroup up the widget's tree.
- Registers the current context with the changes in the FormGroup so that if the validity of the FormGroup changes then the current context is rebuilt.
ReactiveFormConsumer widget is a wrapped around the ReactiveForm.of(context) expression so that we can reimplement the previous example as follows:
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: Column(
children: <Widget>[
ReactiveTextField(
formControlName: 'email',
),
ReactiveTextField(
formControlName: 'password',
obscureText: true,
),
ReactiveFormConsumer(
builder: (context, form, child) {
return RaisedButton(
child: Text('Submit'),
onPressed: form.valid ? _onSubmit : null,
);
},
),
],
),
);
}
void _onSubmit() {
print('Hello Reactive Forms!!!');
}It is entirely up to you to decide which of the above two approaches to use, but note that to access the FormGroup via ReactiveForm.of(context) the consumer widget must always be down in the tree of the ReactiveForm widget.
There are some cases where we want to add or remove focus on a UI TextField without the interaction of the user. For that particular cases you can use FormControl.focus() or FormControl.unfocus() methods.
final form = fb.group({'name': 'John Doe'});
FormControl control = form.control('name');
control.focus(); // UI text field get focus and the device keyboard pop up
control.unfocus(); // UI text field lose focusYou can also set focus directly from the Form like:
final form = fb.group({'name': ''});
form.focus('name'); // UI text field get focus and the device keyboard pop upfinal form = fb.group({
'person': fb.group({
'name': '',
}),
});
// set focus to a nested control
form.focus('person.name');Another example is when you have a form with several text fields and each time the user completes edition in one field you want to request next focus field using the keyboard actions:
final form = fb.group({
'name': ['', Validators.required],
'email': ['', Validators.required, Validators.email],
'password': ['', Validators.required],
});@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: Column(
children: <Widget>[
ReactiveTextField(
formControlName: 'name',
textInputAction: TextInputAction.next,
onSubmitted: () => this.form.focus('email'),
),
ReactiveTextField(
formControlName: 'email',
textInputAction: TextInputAction.next,
onSubmitted: () => this.form.focus('password'),
),
ReactiveTextField(
formControlName: 'password',
obscureText: true,
),
],
),
);
}When you remove focus of a control, the control is marked as touched, that means that the validation error messages will show up in UI. To prevent validation messages to show up you can optionally set argument touched to false.
// remove the focus to the control and marks it as untouched. this.form.unfocus(touched: false);
To disabled a widget like ReactiveTextField all you need to do is to mark the control as disabled:
final form = FormGroup({
'name': FormControl<String>(),
});
FormControl control = form.control('name');
// the control is disabled and also the widget in UI is disabled.
control.markAsDisabled();When a control is disabled it is exempt from validation checks and excluded from the aggregate value of any parent. Its status is DISABLED.
To retrieves all values of a FormGroup or FormArray regardless of disabled status in children use FormControl.rawValue or FormArray.rawValue respectively.
How does ReactiveTextField differs from native TextFormField or TextField?
ReactiveTextField has more in common with TextFormField that with TextField. As we all know TextFormField is a wrapper around the TextField widget that brings some extra capabilities such as Form validations with properties like autovalidate and validator. In the same way ReactiveTextField is a wrapper around TextField that handle the features of validations in a own different way.
ReactiveTextField has all the properties that you can find in a common TextField, it can be customizable as much as you want just as a simple TextField or a TextFormField. In fact must of the code was taken from the original TextFormField and ported to have a reactive behavior that binds itself to a FormControl in a two-way binding.
Below is an example of how to create some ReactiveTextField with some common properties:
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: Column(
children: <Widget>[
ReactiveTextField(
formControlName: 'name',
decoration: InputDecoration(
labelText: 'Name',
),
textCapitalization: TextCapitalization.words,
textAlign: TextAlign.center,
style: TextStyle(backgroundColor: Colors.white),
),
ReactiveTextField(
formControlName: 'phoneNumber',
decoration: InputDecoration(
labelText: 'Phone number',
),
keyboardType: TextInputType.number,
),
ReactiveTextField(
formControlName: 'password',
obscureText: true,
decoration: InputDecoration(
labelText: 'Password',
),
),
],
),
);
}Because of the two-binding capability of the ReactiveTextField with a FormControl the widget don't include properties as controller, validator, autovalidate, onSaved, the FormControl is responsible for handling validation as well as changes notifications.
It does include some events like onChanged, onTab, onEditingComplete, and onSubmitted.
- ReactiveTextField
- ReactiveDropdownField
- ReactiveSwitch
- ReactiveCheckbox
- ReactiveRadio
- ReactiveSlider
- ReactiveCheckboxListTile
- ReactiveSwitchListTile
- ReactiveRadioListTile
- ReactiveDatePicker
- ReactiveTimePicker
- ReactiveForm
- ReactiveFormConsumer
- ReactiveFormBuilder
- ReactiveFormArray
- ReactiveValueListenableBuilder
- ReactiveStatusListenableBuilder
We are trying to keep reactive_forms from bloating with third party dependencies this is why there is
a separate library reactive_forms_widgets which is under construction yet that provides
a variety of more advanced field widgets. To know more about how to install it please visit the library repo and read the documentation about the widgets it contains.
- ReactiveAdvancedSwitch - wrapper around
flutter_advanced_switch - ReactiveDateRangePicker - wrapper around showDateRangePicker
- ReactiveDateTimePicker - wrapper around showDatePicker and showTimePicker
- ReactiveDropdownSearch - wrapper around
dropdown_search - ReactiveFilePicker - wrapper around
file_picker - ReactiveImagePicker - wrapper around
image_picker - ReactiveMultiImagePicker - wrapper around
multi_image_picker - ReactiveSegmentedControl - wrapper around
CupertinoSegmentedControl - ReactiveSignature - wrapper around
signature - ReactiveTouchSpin - wrapper around
flutter_touch_spin - ReactiveRangeSlider - wrapper around
RangeSlider - ReactiveSleekCircularSlider - wrapper around
sleek_circular_slider - ReactiveCupertinoTextField - wrapper around
CupertinoTextField - ReactiveRatingBar - wrapper around
flutter_rating_bar - ReactiveMacosUi - wrapper around
macos_ui - ReactivePinPut - wrapper around
pinput - ReactiveCupertinoSwitch - wrapper around
CupertinoSwitch - ReactivePinCodeTextField - wrapper around
pin_code_fields - ReactiveSlidingSegmentedControl - wrapper around
CupertinoSlidingSegmentedControl - ReactiveCupertinoSlider - wrapper around
CupertinoSlider - ReactiveColorPicker - wrapper around
flutter_colorpicker - ReactiveMonthPickerDialog - wrapper around
month_picker_dialog - ReactiveRawAutocomplete - wrapper around
RawAutocomplete - ReactiveFlutterTypeahead - wrapper around
flutter_typeahead - ReactivePinInputTextField - wrapper around
pin_input_text_field - ReactiveDirectSelect - wrapper around
direct_select - ReactiveMarkdownEditableTextInput - wrapper around
markdown_editable_textinput - ReactiveCodeTextField - wrapper around
code_text_field - ReactivePhoneFormField - wrapper around
phone_form_field - ReactiveExtendedTextField - wrapper around
extended_text_field - ReactiveCupertinoSlidingSegmentedControl - wrapper around
CupertinoSlidingSegmentedControl
We have explain the common usage of a ReactiveTextField along this documentation.
ReactiveDropdownField as all the other reactive field widgets is almost the same as its native version DropdownButtonFormField but adding two-binding capabilities. The code is ported from the original native implementation. It have all the capability of styles and themes of the native version.
final form = FormGroup({
'payment': FormControl<int>(validators: [Validators.required]),
});
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: Column(
children: <Widget>[
ReactiveDropdownField<int>(
formControlName: 'payment',
hint: Text('Select payment...'),
items: [
DropdownMenuItem(
value: 0,
child: Text('Free'),
),
DropdownMenuItem(
value: 1,
child: Text('Visa'),
),
DropdownMenuItem(
value: 2,
child: Text('Mastercard'),
),
DropdownMenuItem(
value: 3,
child: Text('PayPal'),
),
],
),
],
),
);
}As you can see from the above example the usage of ReactiveDropdownField is almost the same as the usage of a common DropdownButtonFormField, except for the additional formControlName and validationMessages properties.
If you want to rebuild a widget each time a FormControl value changes you could use the ReactiveValueListenableBuilder widget.
In the following example we are listening for changes in lightIntensity. We change that value with a ReactiveSlider and show all the time the value in a Text widget:
final form = FormGroup({
'lightIntensity': FormControl<double>(value: 50.0),
});
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: Column(
children: <Widget>[
ReactiveValueListenableBuilder<double>(
formControlName: 'lightIntensity',
builder: (context, value, child) {
return Text('lights at ${value?.toStringAsFixed(2)}%');
},
),
ReactiveSlider(
formControlName: 'lightIntensity',
max: 100.0,
),
],
)
);
}Both widgets are responsible for exposing the FormGroup to descendants widgets in the tree. Let see an example:
// using ReactiveForm
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: ReactiveTextField(
formControlName: 'email',
),
);
}// using ReactiveFormBuilder
@override
Widget build(BuildContext context) {
return ReactiveFormBuilder(
form: () => this.form,
builder: (context, form, child) {
return ReactiveTextField(
formControlName: 'email',
);
},
);
}The main differences are that ReactiveForm is a StatelessWidget so it doesn't save the instance of the FormGroup. You must declare the instance of the FormGroup in a StatefulWidget or resolve it from some Provider (state management library).
// Using ReactiveForm in a StatelessWidget and resolve the FormGroup from a provider
class SignInForm extends StatelessWidget {
@override
Widget build(BuildContext context) {
final viewModel = Provider.of<SignInViewModel>(context, listen: false);
return ReactiveForm(
formGroup: viewModel.form,
child: ReactiveTextField(
formControlName: 'email',
),
);
}
}// Using ReactiveForm in a StatefulWidget and declaring FormGroup in the state.
class SignInForm extends StatefulWidget {
@override
_SignInFormState createState() => _SignInFormState();
}
class _SignInFormState extends State<SignInForm> {
final form = fb.group({
'email': Validators.email,
});
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: this.form,
child: ReactiveTextField(
formControlName: 'email',
),
);
}
}If you declare a FormGroup in a StatelessWidget the group will be destroyed a created each time the instance of the StatelessWidget is destroyed and created, so you must preserve the FormGroup in a state or in a Bloc/Provider/etc.
By the other hand ReactiveFormBuilder is implemented as a StatefulWidget so it holds the created FormGroup in its state. That way is safe to declares the FormGroup in a StatelessWidget or get it from a Bloc/Provider/etc.
class SignInForm extends StatelessWidget {
@override
Widget build(BuildContext context) {
return ReactiveFormBuilder(
form: () => fb.group({'email': Validators.email}),
builder: (context, form, child) {
return ReactiveTextField(
formControlName: 'email',
);
},
);
}
}You should use ReactiveForm if:
- The form is complex enough.
- You need to listen for changes in some child control to execute some business logic.
- You are using some State Management library like Provider or Bloc.
- Using a StatefulWidget to declare a very simple form is something that really doesn't bother you.
You should use ReactiveFormBuilder if:
- The form is quite simple enough and doesn't need a separate Provider/Bloc state.
- You don't want to use a StatefulWidget to declare the FormGroup.
But the final decision is really up to you, you can use any of them in any situations ;)
note: mark your fields with Key's for easy access via widget tester
class LoginForm extends StatefulWidget {
const LoginForm({Key? key}) : super(key: key);
@override
LoginFormState createState() => LoginFormState();
}
class LoginFormState extends State<LoginForm> {
final form = FormGroup({
'email': FormControl<String>(validators: [Validators.required, Validators.email]),
'password': FormControl<String>(validators: [Validators.required]),
});
@override
Widget build(BuildContext context) {
return ReactiveForm(
formGroup: form,
child: Column(
children: <Widget>[
ReactiveTextField(
key: const Key('email'),
formControlName: 'email',
),
ReactiveTextField(
key: const Key('password'),
formControlName: 'password',
obscureText: true,
),
ElevatedButton(
key: const Key('submit'),
onPressed: () {},
child: const Text('Submit'),
),
],
),
);
}
}void main() {
testWidgets('LoginForm should pass with correct values', (tester) async {
// Build the widget.
await tester.pumpWidget(const MaterialApp(
home: Scaffold(body: LoginForm()),
));
await tester.enterText(find.byKey(const Key('email')), '[email protected]');
await tester.enterText(find.byKey(const Key('password')), 'password');
await tester.tap(find.byKey(const Key('submit')));
await tester.pump();
// Expect to find the item on screen if needed
expect(find.text('[email protected]'), findsOneWidget);
// Get form state
final LoginFormState loginFormState = tester.state(find.byType(LoginForm));
// Check form state
expect(loginFormState.form.valid, true);
});
}
Reactive Forms + Provider plugin 💪
Although Reactive Forms can be used with any state management library or even without any one at all, Reactive Forms gets its maximum potential when is used in combination with a state management library like the Provider plugin.
This way you can separate UI logic from business logic and you can define the FormGroup inside a business logic class and then exposes that class to widgets with mechanism like the one Provider plugin brings.
Reactive Forms + code generation 🤖
ReactiveFormsGenerator is the code generator for reactive_forms which will save you tons of time and make your forms type safe.
There is no reason write code manually! Let the code generation work for you.
Reactive Forms is not limited just to common widgets in Forms like text, dropdowns, sliders switch fields and etc, you can easily create custom widgets that two-way binds to FormControls and create your own set of Reactive Widgets ;)
In our Wiki you can find a tutorial of how to create your custom Reactive Widget.
You can also check Star Rating with Flutter Reactive Forms post as another example of a custom reactive widget.
-
Reactive Forms is not a fancy widgets package. It is not a library that brings some new Widgets with new shapes, colors or animations. It lets you to decide the shapes, colors, and animations you want for your widgets, but frees you from the responsibility of gathering and validating the data. And keeps the data in sync between your model and your widgets.
-
Reactive Forms does not pretend to replace the native widgets that you commonly use in your Flutter projects like TextFormField, DropdownButtonFormField or CheckboxListTile. Instead of that it brings new two-way binding capabilities and much more features to those same widgets.
- Reactive Forms provides a model-driven approach to handling form inputs whose values change over time. It's heavily inspired in Angular Reactive Form.
- It lets you focus on business logic and save you time from collect, validate and mantain synchronization between your models and widgets.
- Remove boilerplate code and brings you the posibility to write clean code defining a separation between model and UI with minimal efforts.
- And it integrates perfectly well with common state management libraries like Provider, Bloc and many others good libraries the community has created.
Visit Migration Guide to see more details about different version breaking changes.