com.atlassian.crowd.util.persistence.hibernate.batch
Class BatchProcessorImpl

java.lang.Object
  com.atlassian.crowd.util.persistence.hibernate.batch.BatchProcessorImpl

All Implemented Interfaces:

BatchProcessor

public class BatchProcessorImpl
extends java.lang.Object
implements BatchProcessor

Threadsafe batch processor. Currently capable of performing "find", "saveOrUpdate" and "replicate" on a collection of entities. This processor is essentially a heavyweight generic DAO for processing batching inserts or updates to NamedEntities and for performing finds on DirectoryEntities. A NamedEntity is just an Object with a getName() method suitable for logging the names of objects that could not be processed. It is assumed that all data-cleansing (eg. convert to lower case, directoryID assigned, etc) is performed prior to calling methods on the BatchProcessor. The batchSize defaults to 20 and can be manually set (via Spring, for example) using the appropriate setter method. The batchSize should match the hibernate.jdbc.batch_size property defined in the Hibernate configuration. You can use the BatchProcessor in a singleton manner. There is no need to transaction/Hibernate wrap the BatchProcessor as transactions and Hibernate sessions are managed internally by this class. Each batch operation is first divided into smaller sets of batchSize. Each batchSet is added via a batched JDBC call (using Hibernate's batching mechanism). If there is an error in processing the batch, the batched JDBC call is rolled-back and the batchSet is processed individually. This mechanism ensures very fast (JDBC-batched) inserts and updates and follows it up with a fail-over retry for the failing batches. NOTE 1: do not use this if you're database is not transactional, *stab* MySQL ISAM. NOTE 2: if you are experiencing problems with the BatchProcessor set the log level on this class to DEBUG via log4j.properties: log4j.logger.com.atlassian.crowd.util.persistence.hibernate.BatchProcessor=DEBUG NOTE 3: it is quite is easy to generify this batch processor if you would like batch processing outside Crowd.

Author:

Shihab Hamid

Constructor Summary

BatchProcessorImpl(org.hibernate.SessionFactory sessionFactory)

Method Summary

<E extends DirectoryEntity> java.util.Collection<E>	find(long directoryID, java.util.Collection<java.lang.String> names, java.lang.Class<E> persistentClass) Returns a collection of entities that match the names provided.
<E extends DirectoryEntity> java.util.Collection<E>	processBatchFind(org.hibernate.Session session, long directoryID, java.util.Collection<java.lang.String> names, java.lang.Class<E> persistentClass)
<E extends NamedEntity> BatchResult<E>	replicate(java.util.Collection<E> objects, org.hibernate.ReplicationMode replicationMode) Replicate a set of entities using Hibernate/JDBC batching.
<E extends NamedEntity> BatchResult<E>	saveOrUpdate(java.util.Collection<E> objects) SaveOrUpdate a set of entities using Hibernate/JDBC batching.
void	setBatchSize(int batchSize) The batchSize value should be the same as the hibernate.jdbc.batch_size Hibernate property.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

BatchProcessorImpl

public BatchProcessorImpl(org.hibernate.SessionFactory sessionFactory)

Method Detail

replicate

public <E extends NamedEntity> BatchResult<E> replicate(java.util.Collection<E> objects,
                                                        org.hibernate.ReplicationMode replicationMode)

Description copied from interface: BatchProcessor

Replicate a set of entities using Hibernate/JDBC batching. If some entities fail replication, the rest can still succeed.

Specified by:

replicate in interface BatchProcessor

Parameters:

objects - entities to replicate.

replicationMode - how to perform the replication, eg. ReplicationMode.OVERWRITE.

Returns:

a BatchResult indicating the number of total entities that were attempted to be processed and the collection of entities that failed to be processed.

saveOrUpdate

public <E extends NamedEntity> BatchResult<E> saveOrUpdate(java.util.Collection<E> objects)

Description copied from interface: BatchProcessor

SaveOrUpdate a set of entities using Hibernate/JDBC batching. If some entities fail saving/updating, the rest can still succeed.

Specified by:

saveOrUpdate in interface BatchProcessor

Parameters:

objects - entities to replicate.

Returns:

a BatchResult indicating the number of total entities that were attempted to be processed and the collection of entities that failed to be processed.

find

public <E extends DirectoryEntity> java.util.Collection<E> find(long directoryID,
                                                                java.util.Collection<java.lang.String> names,
                                                                java.lang.Class<E> persistentClass)

Returns a collection of entities that match the names provided. Any names that cannot be matched to persistent entities are not present in the resultant collection. Internally, this performs a: SELECT * FROM entityTable WHERE entityName IN (...) This is batched such that the size of the IN clause is at most the batchSize.

Specified by:

find in interface BatchProcessor

Parameters:

directoryID - directory ID of the entities to return.

names - collection of entity names. This, along with the directoryID should form the primary key of the entity.

persistentClass - the persistent class to lookup. This must be a Hibernate-mapped DirectoryEntity.

Returns:

a collection of the DirectoryEntities that exist matching any of the supplied names.

processBatchFind

public <E extends DirectoryEntity> java.util.Collection<E> processBatchFind(org.hibernate.Session session,
                                                                            long directoryID,
                                                                            java.util.Collection<java.lang.String> names,
                                                                            java.lang.Class<E> persistentClass)

setBatchSize

public void setBatchSize(int batchSize)

The batchSize value should be the same as the hibernate.jdbc.batch_size Hibernate property.

Parameters:

batchSize - batch size used to group batches.