Salesforce Platform Developer Practice Exam

• A. Apex classes and triggers
• B. Objects, fields and relationships Correct answer
• C. Lightning web components
• D. Lightning pages and page layouts

Option B is correct because in MVC on the Salesforce Platform the Model is the data layer: objects, fields and relationships. Apex classes and triggers (A) are the Controller, while Lightning web components (C) and Lightning pages and page layouts (D) are the View.

• A. Org-wide defaults
• B. The role hierarchy
• C. Field-level security
• D. Governor limits Correct answer

Option D is correct because governor limits cap the resources (queries, DML, CPU time, heap) that any one transaction can use on the shared multitenant platform, so runaway code fails instead of starving other tenants. Org-wide defaults, the role hierarchy and field-level security (A, B, C) control data access, not resource consumption.

• A. A Lightning Out app
• B. A Lightning web component Correct answer
• C. A Visualforce page
• D. An Aura component

Option B is correct because Lightning web components are built on native Web Components standards (custom elements, shadow DOM, ES modules). Aura components (D) use a proprietary framework, Visualforce (C) is a server-rendered page technology, and Lightning Out (A) is a way to host components in external apps, not a component model.

• A. The Account object's field definitions
• B. The query in the Apex controller class Correct answer
• C. The page layout assigned to the user's profile
• D. The Visualforce page markup

Option B is correct because which records are returned and how they are sorted is decided by the SOQL query in the controller (the C in MVC). The object's field definitions (A) are the Model and do not change, and the page markup (D) and page layout (C) only affect how data is displayed.

• A. Governor limits keep one tenant's code from using shared resources Correct answer
• B. Developers add database indexes with SQL DDL on their org's tables
• C. The query optimizer relies on tenant-specific data statistics Correct answer
• D. Each org runs Apex in its own virtual machine with private CPU and memory
• E. Apex needs minimum test coverage before it deploys to production Correct answer

Options A, C and E are correct because governor limits, tenant-aware query optimization and the 75% Apex test-coverage requirement for production deployments are all direct consequences of many orgs sharing one platform. Option B is wrong because developers cannot run SQL DDL on the underlying tables, and option D is wrong because orgs share the runtime rather than getting a private virtual machine.

• A. An Apex trigger on Invoice__c that updates Account
• B. A cross-object formula field on Account
• C. An after-save record-triggered flow on Invoice__c that updates Account
• D. A roll-up summary field on Account Correct answer

Option D is correct because a roll-up summary field natively sums Amount__c across the detail records of a master-detail relationship, with no code or automation to maintain. A trigger (A) or flow (C) would work but must be built and maintained, and a cross-object formula (B) can only look up from child to parent, not aggregate children.

• A. final Correct answer
• B. virtual
• C. transient
• D. static

Option A is correct because a final variable can be assigned only once, at declaration or in a constructor/initializer, and any later reassignment is a compile error. virtual applies to classes and methods, transient excludes a field from view state serialization, and static only ties the variable to the class.

• A. Implement Iterable<Deal> and return the deals in order
• B. Implement Comparable and define compareTo(Object other) Correct answer
• C. Override equals and hashCode to compare the amount
• D. Be marked virtual and override the sort() method

Option B is correct because List.sort() orders custom objects only if their class implements the Comparable interface and defines compareTo(Object), where Deal can compare the amount values. Iterable (A) is about looping, equals/hashCode (C) are for equality and Set/Map keys, and sort() cannot be overridden (D).

• A. 3
• B. 4 Correct answer
• C. 8
• D. 10

Option B is correct. The loop adds 0 and 1, skips i = 2 with continue, adds 3, then breaks at i = 4 before adding it, so total = 0 + 1 + 3 = 4. The values 3, 8 and 10 come from miscounting the continue or the break.

• A. It can return the MIN of a Currency field on the invoices Correct answer
• B. It can return the AVG of a Number field on the invoices
• C. It can be limited to invoices that meet filter criteria Correct answer
• D. It can be created on Invoice__c to summarize its Account

Options A and C are correct because a roll-up summary field supports COUNT, SUM, MIN and MAX, so MIN over a Currency field works, and it can be restricted with filter criteria so only matching detail records are included. Option B is wrong because AVG is not an available roll-up type, and option D is wrong because a roll-up summary is defined on the master (Account), never on the detail (Invoice__c).

• A. Cross-site request forgery
• B. Session fixation
• C. Cross-site scripting (XSS) Correct answer
• D. SOQL injection

Option C is correct because escape="false" tells Visualforce to output the field without HTML-encoding it, so any script a user typed into the feedback runs in other users' browsers, which is cross-site scripting. CSRF (A) is mitigated by Visualforce's built-in form tokens, session fixation (B) is unrelated to output encoding, and SOQL injection (D) concerns queries built from input, not page rendering.

• A. lightning__FlowScreen Correct answer
• B. lightning__Tab
• C. lightning__RecordPage
• D. lightning__AppPage

Option A is correct because a component appears as a screen element in Flow Builder only when its js-meta.xml lists the lightning__FlowScreen target (with isExposed set to true). lightning__Tab, lightning__RecordPage and lightning__AppPage place a component on tabs, record pages and app pages, not in flow screens.

• A. this.template.querySelector('lightning-card').showToast('Saved', 'success');
• B. this.dispatchEvent(new ShowToastEvent({ title: 'Saved', variant: 'success' })); Correct answer
• C. this.dispatchEvent(new CustomEvent('toast', { detail: { title: 'Saved', variant: 'success' } }));
• D. ShowToastEvent.fire({ title: 'Saved', variant: 'success' });

Option B is correct because an LWC shows a toast by dispatching a ShowToastEvent imported from lightning/platformShowToastEvent; the variant 'success' gives the green style. lightning-card has no showToast method (A), a generic CustomEvent named 'toast' is not handled by the platform (C), and ShowToastEvent has no static fire() method (D).

• A. ApexPages.currentPage().getParameters().put('error', 'Name is required');
• B. throw new AuraHandledException('Name is required');
• C. ApexPages.addMessage(new ApexPages.Message(ApexPages.Severity.ERROR, 'Name is required')); Correct answer
• D. System.debug(LoggingLevel.ERROR, 'Name is required');

Option C is correct because ApexPages.addMessage adds an ApexPages.Message to the page's message collection, which apex:pageMessages renders. Putting a value in the URL parameters (A) shows nothing, AuraHandledException (B) is for Aura/LWC controllers, and System.debug (D) only writes to the debug log.

• A. A method in a without sharing class ignores sharing rules even when a with sharing class calls it Correct answer
• B. An inner class does not inherit the sharing setting of its outer class Correct answer
• C. Declaring a class with sharing also enforces the running user's object permissions and field-level security
• D. Sharing rules apply to DML in a with sharing class but not to SOQL queries

Options A and B are correct. Sharing is declared per class, so a method in a without sharing class runs without sharing rules even when a with sharing class calls it, and an inner class does not inherit the sharing setting of its outer class. Option C is wrong because with sharing enforces record-level sharing only, not object permissions or field-level security. Option D is wrong because with sharing applies to SOQL as well as DML.

• A. 0 Correct answer
• B. 200
• C. 1,000
• D. 5,000

Option A is correct because a test class annotated with plain @IsTest cannot see org data; it only sees records the test itself creates, so with no setup code the Account count is 0. The 5,000 existing accounts (option D) are visible only with SeeAllData=true, and 200 and 1,000 are not values this query could return.

• A. System.runAs(u) { Assert.areEqual(0, [SELECT COUNT() FROM Account WHERE Id = :acct.Id]); } Correct answer
• B. UserInfo.setUser(u); Assert.areEqual(0, [SELECT COUNT() FROM Account WHERE Id = :acct.Id]);
• C. System.runAs(u.Id) { Assert.areEqual(0, [SELECT COUNT() FROM Account WHERE Id = :acct.Id]); }
• D. Test.runAs(u) { Assert.areEqual(0, [SELECT COUNT() FROM Account WHERE Id = :acct.Id]); }

Option A is correct because System.runAs(User) runs the enclosed block with that user's sharing and permissions, so the query shows what u can see. UserInfo.setUser does not exist as a way to switch context, System.runAs takes a User record rather than an Id, and there is no Test.runAs method.

• A. Call Test.setMock once per endpoint, each time with its own StaticResourceCalloutMock
• B. Register a MultiStaticResourceCalloutMock with one static resource per endpoint and call setStatusCode(404) on it
• C. Implement HttpCalloutMock and pick the response from req.getEndpoint() in respond() Correct answer
• D. Annotate the test with @IsTest(SeeAllData=true) so the callouts reach the real endpoints

Option C is correct because a single HttpCalloutMock implementation can inspect req.getEndpoint() in respond() and return a 200 JSON response for one URL and a 404 for the other. Only one mock is active per test context, so a second Test.setMock replaces the first (option A), and a MultiStaticResourceCalloutMock applies one status code to every response (option B). Tests cannot make real callouts, and SeeAllData does not change that (option D).

• A. Setup records can be queried by the first method that runs, and by no other.
• B. The setup method runs once, not before each test method. Correct answer
• C. Every test method starts with the 50 setup Accounts. Correct answer
• D. Accounts deleted in the first method stay deleted for the other methods.

Options B and C are correct because a @testSetup method runs once for the class, and every test method then starts with its own fresh copy of that data, since each method's changes are rolled back. Option A is wrong because every test method can query the setup records, and option D is wrong because the first method's deletes are rolled back and do not carry over to the other methods.

• A. getMessage() returns 'Script-thrown exception'. Correct answer
• B. getMessage() returns 'Quota exceeded' in the test.
• C. Calling setMessage('Quota exceeded') on the exception before throwing it fixes the text. Correct answer
• D. getMessage() returns the text once the test calls Test.stopTest().

Options A and C are correct because an AuraHandledException built with the constructor message returns the generic 'Script-thrown exception' from getMessage() unless you call setMessage() on it before throwing. Option B is wrong for that reason, and option D is wrong because Test.stopTest() has no effect on the exception message.