Core Data Optimization Gagan Vishal Mishra
Memory Vs. Performance Core Data Optimization 2 Memory Speed Less More Slow Faster
Memory Vs. Performance Core Data Optimization 3 Memory Speed Less More Slow Faster
Thread Safety In Core Data Core Data Optimization 4 • While working with Core Data, it's important to always remember that Core Data isn't thread safe. Core Data expects to be run on a single thread. • NSManagedObject, NSManagedObjectContext NSPersistentStoreCoordinator aren't thread safe. • This doesn't mean that every Core Data operation needs to be performed on the main thread, which is used for UI, but it means that we need to take care which operations are executed on which threads. We have to be careful how changes from one thread are propagated to other threads.
Thread Safety In Core Data Core Data Optimization 5 1. NSManagedObject: NSManagedObject isn't thread safe but has an instance objectID that returns an instance of the NSManagedObjectID class which is thread safe. For getting do { let managedObject = try managedObjectContext.existingObjectWithID(objectID) } catch let fetchError as NSError { print("(fetchError), (fetchError.userInfo)”) } 2. NSManagedObjectContext: Isn't thread safe, we can create a managed object context for every thread that interacts with Core Data. This strategy is referred as thread confinement. A common approach is to create thread dictionary to store MOC. let currentThread = NSThread.currentThread() currentThread.threadDictionary.setObject(managedObjectContext, forKey: "managedObjectContext") 3. NSPersistentStoreCoordinator: We can create a separate persistent store coordinator for every thread. Which is also possible.
Multithreading & Concurrency Strategies Core Data Optimization 6 The NSPersistentStoreCoordinator class was designed to support multiple managed object contexts, even if those managed object contexts were created on different threads. Because the NSManagedObjectContext class locks the persistent store coordinator while accessing it, it is possible for multiple managed object contexts to use the same persistent store coordinator even if those managed object contexts live on different threads. This makes a multithreaded Core Data setup much more manageable and less complex. There are two ways for Core Data Concurrency 1.Notifications 2.Parent/Child Managed Object Contexts Another option is 1.Independent Persistent Store Built with Notification & PrivateQueue
Multithreading & Concurrency Strategies Core Data Optimization 7 There are three types of Concurrency options provided by Apple 1.MainQueueConcurrencyType: The managed object context is only accessible from the main thread. An exception is thrown if you try to access it from any other thread. 1.PrivateQueueConcurrencyType: When creating a managed object context with a concurrency type of PrivateQueueConcurrencyType, the managed object context is associated with a private queue and it can only be accessed from that private queue. 2.ConfinementConcurrencyType: Apple has deprecated this concurrency type as of iOS 9. This is the concurrency type that corresponds with the thread confinement concept . If you create a managed object context using init(), the concurrency type of that managed object context is ConfinementConcurrencyType.
Fetching Core Data Optimization 8 1. Use fetchBatchSize: • This breaks the result set into batches. The entire request will be evaluated, and the identities of all matching objects will be recorded, but no more than batchSize objects' data will be fetched from the persistent store at a time. • A batch size of 0 is treated as infinite, which disables the batch faulting behavior. let fetchRequest = NSFetchRequest(entityName: "DataEntity") fetchRequest.sortDescriptors = [NSSortDescriptor(key: UFO_KEY_COREDATA_SIGHTED, ascending: false)] fetchRequest.fetchBatchSize = 20 //Set number double to visible cells do { let test = try self.managedObjectContext.executeFetchRequest(fetchRequest) as Array print("test count is: (test.count)") } catch let error as NSError { print("Error is : (error.localizedDescription)") }
Fetching Core Data Optimization 9 2. Use resultType: There are four types of result type • ManagedObjectResultType • ManagedObjectIDResultType • DictionaryResultType • CountResultType let fetchRequest = NSFetchRequest(entityName: "DataEntity") fetchRequest.resultType = .DictionaryResultType do { let arrayOfFoundRecords = try objAppDel.tempManagedObjectContext.executeFetchRequest(fetchRequest } catch let error as NSError { print("error in fetch is : (error.localizedDescription)") }
Fetching Core Data Optimization 10 3. Use NSExpressionDescription & NSExpression : Instances of NSExpressionDescription objects represent a special property description type intended for use with the NSFetchRequest PropertiesToFetch method. An NSExpressionDescription describes a column to be returned from a fetch that may not appear directly as an attribute or relationship on an entity. NSExpression is used to represent expressions in a predicate. Comparison operations in an NSPredicate are based on two expressions, as represented by instances of the NSExpression class. Expressions are created for constant values, key paths etc. //Create a expression let expressionDescription = NSExpressionDescription() expressionDescription.name = "count" expressionDescription.expression = NSExpression(forFunction: "count:", arguments: [NSExpression(forKeyPath: "shape")]) //Create a fetch Request let fetchRequest = NSFetchRequest(entityName: "DataEntity")
Fetching Core Data Optimization 11 3. Use GroupBy: Use groupBy to aggregate for same column name. let fetchRequest = NSFetchRequest(entityName: "DataEntity") fetchRequest.propertiesToGroupBy = ["shape"] 4. Use propertiesToFetch: create an array of properties you want to fetch from CoreData. let expressionDescription = NSExpressionDescription() expressionDescription.name = "count” expressionDescription.expression = NSExpression(forFunction: "count:", arguments: [NSExpression(forKeyPath: "shape")]) //Create a fetch Request let fetchRequest = NSFetchRequest(entityName: "DataEntity") fetchRequest.propertiesToFetch = ["shape", expressionDescription]
Fetching Core Data Optimization 12 5. PreFetch Any Required Relationship : Prefetching allows Core Data to obtain developer-specified related objects in a single fetch. let entityDescription = NSEntityDescription.entityForName("Employee", inManagedObjectContext: self.managedObjectContext) let fetchRequest = NSFetchRequest() fetchRequest.entity = entityDescription fetchRequest.relationshipKeyPathsForPrefetching = ["Department"] do{ let arrayOfFoundRecords = try self.managedObjectContext.executeFetchRequest(fetchRequest) as Array for ( emp in arrayOfFoundRecords ) { print("emp is: (emp)") print("dept is: (emp.deptEmp)") } } catch let error as NSError { }
Fetching Core Data Optimization 13 Fetch Summary: 1.Don’t’ fetch anything that you don’t need. 2.Always use Batch Fetch. 3.Try to use NSExpressionDescription for Fetch in other word lets allow SQLite to do calculation. 4.Use Prefeatch (relationshipKeyPathsForPrefetching) for required relationship.
Predicates: Core Data Optimization 14 1. Light Predicate First: Always do String Comparison in last query. For example fetchRequest.predicate = NSPredicate(format: "shape=%@ AND duration=%d", "circle", 30) better way it to avoid string as much as possible or use string as a second argument. So above code will perform better if we write it as: fetchRequest.predicate = NSPredicate(format: "duration=%d shape=%@", 30, "circle”) 2. Use BeginsWith/EndsWith: Always try to use ‘beginsWith’ in creating a predicate.
Predicates: Core Data Optimization 15 Predicate Summary & Cost Graph 1. Do light comparison first i.e. Numerical comparison. 2. Always try to use BeginsWith/EndsWith instead of Equals/Match. 3. Don’t/avoid to use CD in comparison. Predicate Cost Less More BeginsWith/ EndsWith Equality == Contains Matches
Search Core Data Optimization 16 1. Always used Canonical String for search process. For example String Canonical String TEST test åpple apple Green Åpple green apple 2. Always use BeginsWith/EndsWith for searching.
Printing SQL Log Core Data Optimization 17 Product menu by selecting “Manage Schemes” and then editing the current scheme you are using. Select the “Run “phase and then select the “Arguments” tab. Here you will see options to set arguments and environment variables. Add -com.apple.CoreData.SQLDebug 3 to "Arguments Passed On Launch"
18 Thank you !

Core data optimization

  • 1.
  • 2.
    Memory Vs. Performance CoreData Optimization 2 Memory Speed Less More Slow Faster
  • 3.
    Memory Vs. Performance CoreData Optimization 3 Memory Speed Less More Slow Faster
  • 4.
    Thread Safety InCore Data Core Data Optimization 4 • While working with Core Data, it's important to always remember that Core Data isn't thread safe. Core Data expects to be run on a single thread. • NSManagedObject, NSManagedObjectContext NSPersistentStoreCoordinator aren't thread safe. • This doesn't mean that every Core Data operation needs to be performed on the main thread, which is used for UI, but it means that we need to take care which operations are executed on which threads. We have to be careful how changes from one thread are propagated to other threads.
  • 5.
    Thread Safety InCore Data Core Data Optimization 5 1. NSManagedObject: NSManagedObject isn't thread safe but has an instance objectID that returns an instance of the NSManagedObjectID class which is thread safe. For getting do { let managedObject = try managedObjectContext.existingObjectWithID(objectID) } catch let fetchError as NSError { print("(fetchError), (fetchError.userInfo)”) } 2. NSManagedObjectContext: Isn't thread safe, we can create a managed object context for every thread that interacts with Core Data. This strategy is referred as thread confinement. A common approach is to create thread dictionary to store MOC. let currentThread = NSThread.currentThread() currentThread.threadDictionary.setObject(managedObjectContext, forKey: "managedObjectContext") 3. NSPersistentStoreCoordinator: We can create a separate persistent store coordinator for every thread. Which is also possible.
  • 6.
    Multithreading & ConcurrencyStrategies Core Data Optimization 6 The NSPersistentStoreCoordinator class was designed to support multiple managed object contexts, even if those managed object contexts were created on different threads. Because the NSManagedObjectContext class locks the persistent store coordinator while accessing it, it is possible for multiple managed object contexts to use the same persistent store coordinator even if those managed object contexts live on different threads. This makes a multithreaded Core Data setup much more manageable and less complex. There are two ways for Core Data Concurrency 1.Notifications 2.Parent/Child Managed Object Contexts Another option is 1.Independent Persistent Store Built with Notification & PrivateQueue
  • 7.
    Multithreading & ConcurrencyStrategies Core Data Optimization 7 There are three types of Concurrency options provided by Apple 1.MainQueueConcurrencyType: The managed object context is only accessible from the main thread. An exception is thrown if you try to access it from any other thread. 1.PrivateQueueConcurrencyType: When creating a managed object context with a concurrency type of PrivateQueueConcurrencyType, the managed object context is associated with a private queue and it can only be accessed from that private queue. 2.ConfinementConcurrencyType: Apple has deprecated this concurrency type as of iOS 9. This is the concurrency type that corresponds with the thread confinement concept . If you create a managed object context using init(), the concurrency type of that managed object context is ConfinementConcurrencyType.
  • 8.
    Fetching Core Data Optimization 8 1.Use fetchBatchSize: • This breaks the result set into batches. The entire request will be evaluated, and the identities of all matching objects will be recorded, but no more than batchSize objects' data will be fetched from the persistent store at a time. • A batch size of 0 is treated as infinite, which disables the batch faulting behavior. let fetchRequest = NSFetchRequest(entityName: "DataEntity") fetchRequest.sortDescriptors = [NSSortDescriptor(key: UFO_KEY_COREDATA_SIGHTED, ascending: false)] fetchRequest.fetchBatchSize = 20 //Set number double to visible cells do { let test = try self.managedObjectContext.executeFetchRequest(fetchRequest) as Array print("test count is: (test.count)") } catch let error as NSError { print("Error is : (error.localizedDescription)") }
  • 9.
    Fetching Core Data Optimization 9 2.Use resultType: There are four types of result type • ManagedObjectResultType • ManagedObjectIDResultType • DictionaryResultType • CountResultType let fetchRequest = NSFetchRequest(entityName: "DataEntity") fetchRequest.resultType = .DictionaryResultType do { let arrayOfFoundRecords = try objAppDel.tempManagedObjectContext.executeFetchRequest(fetchRequest } catch let error as NSError { print("error in fetch is : (error.localizedDescription)") }
  • 10.
    Fetching Core Data Optimization 10 3.Use NSExpressionDescription & NSExpression : Instances of NSExpressionDescription objects represent a special property description type intended for use with the NSFetchRequest PropertiesToFetch method. An NSExpressionDescription describes a column to be returned from a fetch that may not appear directly as an attribute or relationship on an entity. NSExpression is used to represent expressions in a predicate. Comparison operations in an NSPredicate are based on two expressions, as represented by instances of the NSExpression class. Expressions are created for constant values, key paths etc. //Create a expression let expressionDescription = NSExpressionDescription() expressionDescription.name = "count" expressionDescription.expression = NSExpression(forFunction: "count:", arguments: [NSExpression(forKeyPath: "shape")]) //Create a fetch Request let fetchRequest = NSFetchRequest(entityName: "DataEntity")
  • 11.
    Fetching Core Data Optimization 11 3.Use GroupBy: Use groupBy to aggregate for same column name. let fetchRequest = NSFetchRequest(entityName: "DataEntity") fetchRequest.propertiesToGroupBy = ["shape"] 4. Use propertiesToFetch: create an array of properties you want to fetch from CoreData. let expressionDescription = NSExpressionDescription() expressionDescription.name = "count” expressionDescription.expression = NSExpression(forFunction: "count:", arguments: [NSExpression(forKeyPath: "shape")]) //Create a fetch Request let fetchRequest = NSFetchRequest(entityName: "DataEntity") fetchRequest.propertiesToFetch = ["shape", expressionDescription]
  • 12.
    Fetching Core Data Optimization 12 5.PreFetch Any Required Relationship : Prefetching allows Core Data to obtain developer-specified related objects in a single fetch. let entityDescription = NSEntityDescription.entityForName("Employee", inManagedObjectContext: self.managedObjectContext) let fetchRequest = NSFetchRequest() fetchRequest.entity = entityDescription fetchRequest.relationshipKeyPathsForPrefetching = ["Department"] do{ let arrayOfFoundRecords = try self.managedObjectContext.executeFetchRequest(fetchRequest) as Array for ( emp in arrayOfFoundRecords ) { print("emp is: (emp)") print("dept is: (emp.deptEmp)") } } catch let error as NSError { }
  • 13.
    Fetching Core Data Optimization 13 FetchSummary: 1.Don’t’ fetch anything that you don’t need. 2.Always use Batch Fetch. 3.Try to use NSExpressionDescription for Fetch in other word lets allow SQLite to do calculation. 4.Use Prefeatch (relationshipKeyPathsForPrefetching) for required relationship.
  • 14.
    Predicates: Core Data Optimization 14 1.Light Predicate First: Always do String Comparison in last query. For example fetchRequest.predicate = NSPredicate(format: "shape=%@ AND duration=%d", "circle", 30) better way it to avoid string as much as possible or use string as a second argument. So above code will perform better if we write it as: fetchRequest.predicate = NSPredicate(format: "duration=%d shape=%@", 30, "circle”) 2. Use BeginsWith/EndsWith: Always try to use ‘beginsWith’ in creating a predicate.
  • 15.
    Predicates: Core Data Optimization 15 PredicateSummary & Cost Graph 1. Do light comparison first i.e. Numerical comparison. 2. Always try to use BeginsWith/EndsWith instead of Equals/Match. 3. Don’t/avoid to use CD in comparison. Predicate Cost Less More BeginsWith/ EndsWith Equality == Contains Matches
  • 16.
    Search Core Data Optimization 16 1.Always used Canonical String for search process. For example String Canonical String TEST test åpple apple Green Åpple green apple 2. Always use BeginsWith/EndsWith for searching.
  • 17.
    Printing SQL Log CoreData Optimization 17 Product menu by selecting “Manage Schemes” and then editing the current scheme you are using. Select the “Run “phase and then select the “Arguments” tab. Here you will see options to set arguments and environment variables. Add -com.apple.CoreData.SQLDebug 3 to "Arguments Passed On Launch"
  • 18.

Editor's Notes

  • #10 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/
  • #11 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/
  • #12 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/
  • #13 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/
  • #14 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/
  • #15 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/
  • #16 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/
  • #17 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/
  • #18 http://www.cimgf.com/2015/06/25/core-data-and-aggregate-fetches-in-swift/ https://www.bignerdranch.com/blog/new-in-core-data-and-ios-8-batch-updating/