allow null value persistence and retrieval #111
Conversation
|
|
||
| private static final Set<String> COMPARISON_OPERATOR_NAMES_SUPPORTED = | ||
| Set.of("$eq", "$ne", "$gt", "$gte", "$lt", "$lte"); | ||
|
|
There was a problem hiding this comment.
HIBERNATE-74 exceptions are intertwinced with HIBERNATE-48; we have to refrain from throwing HIBERNATE-74 exceptions from JDBC flow to unblock HIBERNATE-48. For this reason, I decoupled them and created a separate HIBERNATE-74 protected mechanism as above. It is more robust and is free of cache implication.
Given HIBERNATE-74 is to be done soon, we might well simply ignore the above protection.
| if (domainValue == null) { | ||
| throw new FeatureNotSupportedException( | ||
| "TODO-HIBERNATE-48 https://jira.mongodb.org/browse/HIBERNATE-48 return toBsonValue(domainValue)"); | ||
| return null; |
There was a problem hiding this comment.
we have to return null instead of BsonNull for we need to return domain value for bindValue; Struct -> BsonDocument transformation should be the only exception; for that reason, I changed the method return type to better align with its purpose.
| return extractor.getJavaType().wrap(domainObjects, options); | ||
| } else { | ||
| return extractor.getJavaType().wrap(array, options); | ||
| } |
There was a problem hiding this comment.
The above code changes is to bypass a newly found Hibernate ORM bug. Luckily in this case we can work around.
I created the Hibernate ORM ticket at https://hibernate.atlassian.net/browse/HHH-19596. The bugfix has been merged to both v7 and v6.6.
Just a couple of days we could eliminate the above change after a new v6.6 Hibernate ORM version is released.
There was a problem hiding this comment.
I bumped the version to latest v6.6.21 and found the issue has been solved.
https://jira.mongodb.org/browse/HIBERNATE-48
The ticket has both tech design part and implementation part:
The main decisions from design are summarized as below:
The main work in this PR includes:
Some tech decisions in this PR:
@DynamicInsertfor two reasons: 1) it was disabled for protective development reason during previous PR; now the reason ceases to exist; 2) we can't avoid mixture of null and missing fields even if we disabled@DynamicInsertfor HQL partial insertion statement will end up with the mixture easily