When inserting bean with transparent persistence the beans are registered with the persistence context. This is done to support the cases where the beans are inserted, a flush() occurs, the inserted bean is mutated and now dirty - we want that to be detected and get an update
When delete bean executed with transparent persistence, this marks the bean as removed from the persistence context early.
This avoids a "dirty" deleted bean from being seen as a "dirty" bean in the persistence context at flush() time.
collectQueryPlans -> queryPlanEnable
collectQueryPlanThresholdMicros -> queryPlanThresholdMicros
properties keys:
ebean.collectQueryPlans -> ebean.queryPlan.enable
ebean.collectQueryPlanThresholdMicros -> ebean.queryPlan.thresholdMicros
This then aligns all the queryPlan related properties together which is important in order to avoid confusion.
Fix for - standard JPA AttributeConverter which doesn't have the getNullValue method, he/she can't convert null to a custom null object
AttributeConverterAdapter probes the AttributeConverter for the nullValue rather than just assuming it is null.
Currently we can not use @DbName with the entities of the default database. With this change the generated code that registers entity classes will support using @DbName with the default database.
This is a variation of findEach() that makes it easy to have a batch consumer processing a large result in batches. For example, process in batches of 50 beans.
Note that the last batch consumed/processed will often have less than the batch size.
When using nativeSql (think of it as the "root query") we can't use "fetch joins" (FetchConfig.ofDefault()) as that means to prefer to use a SQL JOIN. In this nativeSql case we need to effectively automatically convert fetch() to fetchQuery()
The improvement here is that SqlTreeNodeManyRoot uses load and isContextBean() so that it only checks if the collection already contains the detailBean if the detailBean was already in the persistence context (aka not newly loaded and added). The vast majority case is that detailBean is loaded fresh so for the common case we can skip the extra collection contains() check.
Adds SqlTreeRoot interface to make it more explicit that SqlTreeNodeRoot is the only valid top level node in the sql tree (and as such has the root level load() method)
* Update test only - update TestBasicClobNoVer to use DB and AssertJ
* Refactor OrmQueryProperties internals to reduce raw property parsing
- Remove String properties
- Reduces parsing of properties via fetchProperties() selectProperties() methods
* No effective change - tidy pom.xml for ebean-querybean
* Refactor query internals adding fetchProperties() selectProperties()
- Adds SpiQueryFetch
- Used by query beans with Set<String> properties passed (effectively skipping any parsing)
Robin Bygrave