Salesforce Certified Platform Developer II Free Sample Questions

Using Database.Savepoint and Database.rollback() provides explicit control over the transaction. By setting a savepoint before attempting the DML operations, the developer can catch any exception, roll back all changes to the state of the database before the savepoint was set, and then return a controlled error message. A simple try-catch with a thrown exception also rolls back the transaction, but using a savepoint is the canonical and most explicit way to manage partial transaction rollbacks or full rollbacks while maintaining control over the execution flow to return a specific response.

To make LWC properties configurable in the Lightning App Builder, they must be declared as public properties using the @api decorator in the component's JavaScript file. Additionally, these properties must be exposed in the component's configuration file (.js-meta.xml) within the section. This allows administrators to set values for these properties directly in the App Builder's UI.

For processing large data sets (over 50,000 records), Database.QueryLocator is the optimal choice for the start method of a Batch Apex class. It bypasses the heap size limit for the total number of records retrieved by the query because it fetches records in chunks. In contrast, returning an Iterable or a List loads all records into the heap at once, which would cause a heap limit exception with over 1 million records.

The view state error is caused by large amounts of data being stored in the controller's properties between server requests. The transient keyword prevents a variable from being saved as part of the view state. While this means the data will need to be re-fetched on each postback, it directly solves the 'Maximum view state size' error by removing the large data list from the view state.

Custom Metadata Types are the best fit for this scenario. They are deployable, meaning the configuration data can be moved between environments as metadata. They can be queried using SOQL, and they support relationships, allowing the developer to model the parent-child structure between 'Service Levels' and 'Country Surcharges'. Custom Objects are data, not metadata, and are harder to deploy. Custom Settings are not as well-suited for complex relational structures and are less easily deployed.

The query is timing out because it is non-selective. It has to perform a full table scan on 10 million records to find contacts where LastLoginDate is older than 180 days. Standard date fields like LastLoginDate are not indexed by default. The most effective solution is to work with Salesforce Support to create a custom index on this field. This will make the query highly selective and allow it to execute efficiently without timing out.

This statement is true. The External ID field must uniquely identify a single record for the upsert operation to succeed. If the value of the External ID field is not unique and matches multiple existing records, the operation will fail and a DmlException will be thrown, indicating that multiple records were found for the upsert.

The @InvocableMethod annotation is specifically designed to expose an Apex method as a callable action in declarative tools like Process Builder, Flow, and the REST API. The method must be static and can have only one parameter, which must be a list of a primitive data type or a custom Apex class.

Queueable Apex is the best fit. It runs asynchronously, so it won't delay the Opportunity save. Crucially, a Queueable job can enqueue another job from within its execute method, which allows for complex patterns like implementing a retry mechanism. The job can maintain a retryCount state variable and, upon failure, re-enqueue itself if the count is below the threshold. @future methods cannot call other @future methods, and Batch Apex is overkill for a single callout.