[#BATCH-150] Create HibernateInputSource

Spring Batch

Create HibernateInputSource

Created: 01/Oct/07 10:03 AM Updated: 04/Mar/08 11:39 AM

Component/s:

Infrastructure

Affects Version/s:

1.0-m2

Fix Version/s:

1.0-m3

Security Level:

Public (Public Issues)

Time Tracking:

Not Specified

File Attachments:

hibernate-infrastructure.patch (89 kB)
2.

hibernate-samples.patch (28 kB)

Description

« Hide

Create a HibernateInputSource, keeping in mind Hibernate's recommendations for running in a Batch application:

http://www.hibernate.org/hib_docs/reference/en/html/batch.html

All

Comments

Work Log

Change History

FishEye

Builds

Sort Order:

[ Permlink | « Hide ]

Robert Kasanicky added a comment - 03/Oct/07 07:29 AM

Attached patches contain:

- hibernate input source based on stateless session and iteration over scrollable resultset
- unified tests for all sql input sources (one common superclass contains testcases)
- necessary refactorings to SingleKeyDrivingQuery (new functionality added) and SqlCursor (minor tweaks) input sources
- sample job using hibernate input source for input and hibernate dao for output

[ Permlink | « Hide ]

Dave Syer added a comment - 04/Oct/07 10:20 AM

I'm not so sure a stateless session is the best approach - you lose the benefit of 1st level cache. Needs to be tested against a realistic complex data model (with reference data).

[ Permlink | « Hide ]

Wayne Lund added a comment - 04/Oct/07 06:23 PM

I don't believe there's a lot of benefit to 1st level cache in batch mode with Hibernate. It would be interesting to have a hibernate developer comment. Reference Data is typically stored in 2nd level cache but I wonder if it would get there without the connection through 1st level cache. I agree that in batch you do tend to initialize reference data that will be accessed throughout the batch job. I would be willing to manually populate level 2 cache to have the benefits of bypassing managed objects in batch.

[ Permlink | « Hide ]

Lucas Ward added a comment - 15/Oct/07 12:52 AM

A hibernate input source has been added, using the patch added to ~~BATCH-144~~.

[ Permlink | « Hide ]

Paul Benedict added a comment - 04/Mar/08 11:08 AM

Is this related to the work on SPR-2495?

[ Permlink | « Hide ]

Dave Syer added a comment - 04/Mar/08 11:39 AM

Possibly - except that SPR-2495 is asking about updates and inserts and this is focused on input. The implementation is now called HibernateCursorItemReader if you are interested.