CODEWITHSUNDEEP
javaspringjpatransactions
Stanislav Mikheyev
Stanislav Mikheyev

Reputation: 39

How Spring JPA transactions store data before commit

Lets say I need to create 100.000 records and than .save() them. All records should be saved together and only one transaction is allowed. Clearly, this is a problem and program can run out of memory. But what will happen if I will do the same work in the same transaction but in chunks? 1000 records per iteration and than I save them. Will it fix the problem? Intuitively I think it will not and design is wrong. What happens with entities after I .save()? Does JPA still keep references to all saved entities? Where can I read more on this topic? Thanks.

Upvotes: 0

Views: 501

Answers (1)

crizzis
crizzis

Reputation: 10716

But what will happen if I will do the same work in the same transaction but in chunks? 1000 records per iteration and than I save them. Will it fix the problem?

It will fix the problem if you use the following approach:

while (hasNextBatch()) {
    saveBatch();
    entityManager.flush();
    entityManager.clear();
}

What happens with entities after I .save()? Does JPA still keep references to all saved entities?

Yes (unless you do the above).

Where can I read more on this topic?

Here, for example.

Upvotes: 1

Related Questions