use of com.evolveum.midpoint.repo.sql.query.hqm.condition.Condition in project midpoint by Evolveum.
the class ItemPathResolver method findExistingAlias.
private String findExistingAlias(RootHibernateQuery hibernateQuery, JpaLinkDefinition<?> joinedItemDefinition, String joinedItemFullPath) throws QueryException {
for (JoinSpecification existingJoin : hibernateQuery.getPrimaryEntity().getJoinsFor(joinedItemFullPath)) {
// but let's check the condition as well
String existingAlias = existingJoin.getAlias();
// we have to create condition for existing alias, to be matched to existing condition
Condition conditionForExistingAlias = createJoinCondition(existingAlias, joinedItemDefinition, hibernateQuery);
if (Objects.equals(conditionForExistingAlias, existingJoin.getCondition())) {
LOGGER.trace("Reusing alias '{}' for joined path '{}'", existingAlias, joinedItemFullPath);
return existingAlias;
}
}
return null;
}
use of com.evolveum.midpoint.repo.sql.query.hqm.condition.Condition in project midpoint by Evolveum.
the class ItemPathResolver method reuseOrAddJoin.
String reuseOrAddJoin(JpaLinkDefinition<?> joinedItemDefinition, String currentHqlPath, boolean reuseMultivaluedJoins) throws QueryException {
RootHibernateQuery hibernateQuery = context.getHibernateQuery();
String joinedItemJpaName = joinedItemDefinition.getJpaName();
String joinedItemFullPath = currentHqlPath + "." + joinedItemJpaName;
String joinedItemAlias;
if (joinedItemDefinition.isMultivalued()) {
if (reuseMultivaluedJoins) {
String existingAlias = findExistingAlias(hibernateQuery, joinedItemDefinition, joinedItemFullPath);
if (existingAlias != null) {
return existingAlias;
} else {
// TODO better explanation
throw new QueryException("Couldn't reuse existing join for " + joinedItemDefinition + " in the context of " + currentHqlPath);
}
}
} else {
/*
* Let's check if we were already here i.e. if we had already created this join.
* This is to avoid useless creation of redundant joins for singleton items.
*
* But how can we be sure that item is singleton if we look only at the last segment (which is single-valued)?
* Imagine we are creating join for single-valued entity of u.abc.(...).xyz.ent where
* ent is single-valued and not embedded (so we are to create something like "left join u.abc.(...).xyz.ent e").
* Then "u" is certainly a single value: either the root object, or some value pointed to by Exists restriction.
* Also, abc, ..., xyz are surely single-valued: otherwise they would be connected by a join. So,
* u.abc.(...).xyz.ent is a singleton.
*
* Look at it in other way: if e.g. xyz was multivalued, then we would have something like:
* left join u.abc.(...).uvw.xyz x
* left join x.ent e
* And, therefore we would not be looking for u.abc.(...).xyz.ent.
*/
String existingAlias = findExistingAlias(hibernateQuery, joinedItemDefinition, joinedItemFullPath);
if (existingAlias != null) {
return existingAlias;
}
}
joinedItemAlias = hibernateQuery.createAlias(joinedItemDefinition);
Condition condition = createJoinCondition(joinedItemAlias, joinedItemDefinition, hibernateQuery);
hibernateQuery.getPrimaryEntity().addJoin(new JoinSpecification(joinedItemAlias, joinedItemFullPath, condition));
return joinedItemAlias;
}
Aggregations