Skip to main content

Mocking

Mocking provides a way to substitute records from a database with some prepared data. Data can be prepared in the form of SObject records and lists in Apex code or Static Resource .csv file. Mocked queries won't make any SOQLs and simply return data set in method definition. Mock will ignore all filters and relations; what is returned depends solely on data provided to the method. Mocking works only during test execution. To mock a SOQL query, use the .mockId(id) method to make it identifiable. If you mark more than one query with the same ID, all marked queries will return the same data.

ExampleController.cls
public with sharing class ExampleController {

public static List<Account> getPartnerAccounts(String accountName) {
return SOQL_Account.query()
.with(Account.BillingCity, Account.BillingCountry)
.whereAre(SOQL.FilterGroup
.add(SOQL.Filter.name().contains(accountName))
.add(SOQL.Filter.recordType().equal('Partner'))
)
.mockId('ExampleController.getPartnerAccounts')
.toList();
}
}

Then in tests simply pass data you want to get from the Selector to the SOQL.mock(id).thenReturn(data) method. Acceptable formats are: List<SObject>, SObject, Integer (for count queries), or List<Map<String, Object>> (for aggregate results). During execution, the Selector will return the desired data.

Insights

Mocking Stack Functionality

SOQL Lib implements a sophisticated mocking system that supports multiple sequential mocks for the same query identifier. This enables complex testing scenarios where the same query needs to return different results across multiple executions.

How the Stack Works:

Each call to SOQL.mock(mockId) creates a new mock entry and adds it to a list (stack) associated with that mock ID. When queries are executed:

  • Single Mock: If only one mock exists for the ID, it's reused for all executions
  • Multiple Mocks: Mocks are consumed in FIFO (First In, First Out) order - each execution removes and returns the first mock from the stack
// Setup multiple sequential mocks
SOQL.mock('testQuery').thenReturn(new Account(Name = 'First Call'));
SOQL.mock('testQuery').thenReturn(new Account(Name = 'Second Call'));
SOQL.mock('testQuery').thenReturn(new Account(Name = 'Third Call'));

// First execution returns "First Call", then removes that mock
Account result1 = SOQL.of(Account.SObjectType).mockId('testQuery').toObject();

// Second execution returns "Second Call", then removes that mock
Account result2 = SOQL.of(Account.SObjectType).mockId('testQuery').toObject();

// Third execution returns "Third Call", then removes that mock
Account result3 = SOQL.of(Account.SObjectType).mockId('testQuery').toObject();

// Fourth execution would fail - no more mocks available

Id Field Behavior

The Id field is always included in mocked results, even if it wasn't explicitly specified. This is designed to mirror standard SOQL behavior — Salesforce automatically includes the Id field in every query, even when it's not listed in the SELECT clause.

List<Account> accounts = [SELECT Name FROM Account LIMIT 3];
System.debug(accounts);
/* Output:
(
Account:{Name=Test 1, Id=001J5000008AvzkIAC},
Account:{Name=Test 2, Id=001J5000008AvzlIAC},
Account:{Name=Test 3, Id=001J5000008AvzmIAC}
)
*/

Similarly, when you mock records using SOQL Lib:

SOQL.mock('mockingQuery').thenReturn(new List<Account>{
new Account(Name = 'Test 1'),
new Account(Name = 'Test 2')
});

List<Account> accounts = SOQL.of(Account.SObjectType)
.with(Account.Name)
.mockId('mockingQuery')
.toList();

/* Output:
(
Account:{Name=Test 1, Id=001J5000008AvzkIAC},
Account:{Name=Test 2, Id=001J5000008AvzlIAC}
)
*/

Even though Id wasn’t specified in .with(), it’s automatically added.

Stripping Additional Fields

When using test data factories or builders, it’s common to generate records populated with many fields. However, SOQL Lib ensures that only the fields specified in the query are retained — all other fields are stripped to simulate real SOQL behavior.

// Setup
List<Account> accounts = new List<Account>{
new Account(Name = 'Test 1', Description = 'Test 1 Description', Website = 'www.beyondthecloud.dev'),
new Account(Name = 'Test 2', Description = 'Test 2 Description', Website = 'www.beyondthecloud.dev')
};

// Test
SOQL.mock('mockingQuery').thenReturn(accounts);

List<Account> result = SOQL.of(Account.SObjectType)
.with(Account.Name)
.mockId('mockingQuery')
.toList();

for (Account mockedResult : result) {
Assert.isTrue(mockedResult.isSet(Account.Name), 'Only Account Name should be set.');
Assert.isNull(mockedResult.Description, 'The Account Description should not be set because it was not included in the SELECT clause.');
Assert.isNull(mockedResult.Website, 'The Account Website should not be set because it was not included in the SELECT clause.');
}

In this case:

  • Although Description and Website were present in the mocked data, they are removed because they weren’t part of the query.
  • Only fields explicitly defined in .with() (plus Id by default) remain. Account Description and Website are null, even though they were specified. However, SOQL specified only Account.Name, so additional fields are stripped.

Note: Currently, this field-stripping behavior applies only to simple fields (such as Name, Description, etc.) and subqueries. Additional subqueries are also removed — only the explicitly queried subqueries remain.

Relationship fields and queries using functions are not yet covered by this logic. This means that SOQL queries involving relationship records or functions (like COUNT, AVG) will return the exact mock defined in the unit test. This may be addressed in future enhancements.

Queries Issued Count

Mocked queries in SOQL Lib are counted towards the SOQL query limit, just like real queries. If the number of issued queries exceeds the limit, SOQL Lib will throw:

QueryException: Too many SOQL queries.

This behavior is consistent with Salesforce’s native limits, ensuring that your unit tests accurately reflect potential production scenarios.

List of records

ExampleControllerTest.cls
@IsTest
private class ExampleControllerTest {

@IsTest
static void getPartnerAccounts() {
List<Account> accounts = new List<Account>{
new Account(Name = 'MyAccount 1'),
new Account(Name = 'MyAccount 2')
};

SOQL.mock('ExampleController.getPartnerAccounts').thenReturn(accounts);

// Test
List<Account> result = ExampleController.getPartnerAccounts('MyAccount');

Assert.areEqual(accounts, result);
}
}

Single record

ExampleControllerTest.cls
@IsTest
private class ExampleControllerTest {

@IsTest
static void getPartnerAccount() {
SOQL.mock('ExampleController.getPartnerAccount').thenReturn(new Account(Name = 'MyAccount 1'));

// Test
Account result = (Account) ExampleController.getPartnerAccounts('MyAccount');

Assert.areEqual('MyAccount 1', result.Name);
}
}

Static resource

ExampleControllerTest.cls
@IsTest
private class ExampleControllerTest {

@IsTest
static void getPartnerAccounts() {
SOQL.mock('ExampleController.getPartnerAccounts').thenReturn(Test.loadData(Account.SObjectType, 'ProjectAccounts'));

// Test
List<Account> result = ExampleController.getPartnerAccounts('MyAccount');

Assert.areEqual(5, result.size());
}
}

Count Result

ExampleControllerTest.cls
@IsTest
private class ExampleControllerTest {

@IsTest
static void getPartnerAccountsCount() {
SOQL.mock('mockingQuery').thenReturn(2);

Integer result = SOQL.of(Account.sObjectType).count().mockId('mockingQuery').toInteger();

Assert.areEqual(2, result);
}
}

Sub-Query

To mock a sub-query, we need to use a deserialization mechanism. There are two approaches: using JSON string or Serialization/Deserialization. After deserialization to the desired SObjectType, pass the data to SOQL by calling the .mock method.

Using JSON String

By passing a simple String, it is possible to write non-writable fields, like Name on the Contact object.

@IsTest
static void getAccountsWithContacts() {
List<Account> mocks = (List<Account>) JSON.deserialize(
'[{ "Name": "Account Name", "Contacts": { "totalSize": 1, "done": true, "records": [{ "Name": "Contact Name", "Email": "contact.email@address.com" }] } }],
List<Account>.class
);

SOQL.mock('AccountsController.getAccountsWithContacts').thenReturn(mocks);

List<Account> accounts;

Test.startTest();
accounts = AccountsController.getAccountsWithContacts();
Test.stopTest();

Assert.isNotNull(accounts);
Assert.isNotNull(accounts[0].contacts);
Assert.areEqual(1, accounts[0].contacts.size());
}

Using Serialization/Deserialization

Using this approach, it is possible to bind data with additional logic, like using a Test Data Factory.

@IsTest
static void getAccountsWithContacts() {
List<Account> mocks = (List<Account>) JSON.deserialize(
JSON.serialize(
new List<Map<String, Object>>{
new Map<String, Object>{
'Name' => 'Account Name',
'Contacts' => new Map<String, Object>{
'totalSize' => 1,
'done' => true,
'records' => new List<Contact>{ new Contact(FirstName = 'Contact', LastName = 'Name', Email = 'contact.email@address.com') }
}
}
}
),
List<Account>.class
);

SOQL.mock('AccountsController.getAccountsWithContacts').thenReturn(mocks);

List<Account> accounts;

Test.startTest();
accounts = AccountsController.getAccountsWithContacts();
Test.stopTest();

Assert.isNotNull(accounts);
Assert.isNotNull(accounts[0].contacts);
Assert.areEqual(1, accounts[0].contacts.size());
}

Parent relationship

@IsTest
private class ExampleControllerTest {
@IsTest
static void getPartnerAccountsCount() {
SOQL.mock('mockingQuery').thenReturn(
new Account(
Name = 'Test',
Parent = new Account(Name = 'Parent Name')
)
);

Account result = (Account) ExampleController.getPartnerAccounts('MyAccount');

Assert.areEqual(2, result);
}
}

AggregateResult

There is no way to create a new AggregateResult() instance in Apex. You can find more details here: AggregateResult mocking consideration.

To mock AggregateResult, we introduced SOQL.AggregateResultProxy, which provides the same methods as the standard AggregateResult class.

ExampleController.cls
public with sharing class ExampleController {
public void getLeadAggregateResults() {
List<SOQL.AggregateResultProxy> result = SOQL.of(Lead.SObjectType)
.with(Lead.LeadSource)
.COUNT(Lead.Id, 'total')
.groupBy(Lead.LeadSource)
.mockId('ExampleController.getLeadAggregateResults')
.toAggregatedProxy(); // <== use toAggregatedProxy()
}
}
ExampleControllerTest.cls
@IsTest
public class ExampleControllerTest {
@IsTest
static void getLeadAggregateResults() {
List<Map<String, Object>> aggregateResults = new List<Map<String, Object>>{
new Map<String, Object>{ 'LeadSource' => 'Web', 'total' => 10},
new Map<String, Object>{ 'LeadSource' => 'Phone', 'total' => 5},
new Map<String, Object>{ 'LeadSource' => 'Email', 'total' => 3}
};

SOQL.mock('ExampleController.getLeadAggregateResults').thenReturn(aggregateResults);

Test.startTest();
List<SOQL.AggregateResultProxy> result = ExampleController.getLeadAggregateResults();
Test.stopTest();

// Assert
Assert.areEqual(3, result.size(), 'The size of the aggregate results should match the mocked size.');
}
}

No Results

Pass an empty list: .thenReturn(new List<Type>());

  • When .toList() is invoked, it will return a List<Type>.
  • When .toObject() is invoked, it will return null.

This behavior will be the same as it is during runtime.

ExampleControllerTest.cls
@IsTest
public class ExampleControllerTest {
private static final String TEST_ACCOUNT_NAME = 'MyAccount 1';

@IsTest
static void getAccountByName() {
SOQL.mock('ExampleController.getAccountByName')
.thenReturn(new List<Account>());

Test.startTest();
Account result = (Account) ExampleController.getAccountByName(TEST_ACCOUNT_NAME);
Test.stopTest();

Assert.isNull(result);
}
}