JdbcAggregateOperations delete by query #2084
Conversation
Issue link: spring-projects#1978 Add deleteAllByQuery method to JdbcAggregateOperations This method enables deleting aggregates based on a query by performing the following steps: 1. Select root IDs matching the query with a SELECT ... FOR UPDATE to lock the rows. 2. Delete all sub-entities associated with the selected root IDs. 3. Delete the root entities identified by the selected IDs. Signed-off-by: JaeYeon Kim <[email protected]>
spring-projects-issues
added
the
status: waiting-for-triage
|
I’ve opened a draft PR for #1978. If you have any suggestions for improvement or spot any issues, I’d be glad to make the necessary changes. Thank you for reviewing it! |
mp911de
added
type: enhancement
There was a problem hiding this comment.
Most of this looks pretty good.
I have on major concern though: We should not transfer the ids from the locking select to the client and use it for follow up deletes.
Instead the locking select should execute on the server only, and the delete operations should be based on the query directly.
The reason for this that this operation is very attractive for large amounts of data. In such cases transporting and converting the ids creates serious load on network and memory.
And if someone prefers to select the ids first and then delete based on these ids they can still do that using multiple methods.
Issue link: #1978
Add deleteAllByQuery method to JdbcAggregateOperations
This method enables deleting aggregates based on a query by performing the following steps:
see
#1978