Jpa Repository save inside a for loop

Question

I have this requirement:

• Csv Upload (using opencsv)
• Should display if the record is inserted (if it doesn't exist in the db) or updated (if it exists). Eg.: 5 records inserted 2 records updated.

Usually I use repo.saveAll(entityObjList) to save all the entries.

Since I have to keep track of number of updates and inserts, I came up with using the repo.save(entityObj) method inside the for loop which is not greatly appreciated.

Is there any other way to approach this problem ? Any suggestion is welcome.

• Robert Harvey: What does "which is not greatly appreciated" mean? Did you get some pushback from your team/corporate? What was the feedback?
• Robert Harvey: Most relational database systems will return the number of records affected from a query. Why don't you just fix your saveAll method so it returns that number?
• padmanabhanm: @RobertHarvey Well we decided not to use the JpaRepository.save() inside the for loop because of this. The JpaRepository.saveAll() seems to return the last saved object. It is not my implementation of the same.
• Robert Harvey: If you can't fix the saveAll method, you'll have to decide whether the performance or the number of records returned is more important.
• Delioth: Since you're using Spring: unless you're explicitly building your own JpaRepository implementation, calling saveAll from the default SimpleJpaRepository... just does a for loop and calls save on every entity therein. So doing a for loop and saving each thing... is exactly the same performance as saveAll.

Answer 1

If the count of the inserts/updates is the only thing you need, quick (and possibly dirty) solution might be to keep using saveAll method and selecting the count of the entities in the database before and after the saveAll call -- all wrapped in a transaction of course. Knowing the size of the collection passed to saveAll you can calculate how many entities were created and how many were updated in the database.

Answer 2

Since the Ids of the entities are numeric and generated by the database on insert, loop over all of the entities. If the Id of an entity is 0 or null, then it will be an INSERT. If the Id of an entity is greater than zero (or what ever the minimum number of the sequence is) then it will be an UPDATE.

You would loop over the entities to be saved and tally up in the inserts and updates, then call the saveAll method on the repository, where you can get the performance boost over calling save for each entity individually.

Summing the inserts and updates might be better off in a service layer or class representing a use case, rather than the repository. The additional loop to detect the inserts and updates will add overhead, but you should be fine performance-wise until you reach the 1,000s of entities.

Even then, measure the performance to see if the decrease is small enough to make this feature worthwhile.

Answer 3

From org.springframework.data.jpa.repository.support.SimpleJpaRepository

/*
 * (non-Javadoc)
 * @see org.springframework.data.repository.CrudRepository#save(java.lang.Object)
 */
 @Transactional
 public <S extends T> S save(S entity) {
 if (entityInformation.isNew(entity)) {
 em.persist(entity);
 return entity;
 } else {
 return em.merge(entity);
 }
 }
//...
@Transactional
 public <S extends T> List<S> save(Iterable<S> entities) {
 List<S> result = new ArrayList<S>();
 if (entities == null) {
 return result;
 }
 for (S entity : entities) {
 result.add(save(entity));
 }
 return result;
 }

Basically, you can mimic what save(Iterable) does. Use the annotation @Transactional in your repository method and choose the propagation strategy that best works for you. REQUIRED should be enough.