CODEWITHSUNDEEP
javacouchbaserx-java
panjiesw
panjiesw

Reputation: 156

Couchbase bulk retrieving multiple keys in Java

In couchbase, consider a document has a field which contains a set of keys referencing other documents

{
    "some_ids": ["otherdoc1", "otherdoc2", "otherdoc3"]
}

Which one of these 2 solutions for retrieving all doc in some_ids field gives the best performance?

  1. Batching with RxJava

    List<JsonDocument> foundDocs = Observable
.just("otherdoc1", "otherdoc2", "otherdoc3")
.flatMap(new Func1<String, Observable<JsonDocument>>() {
    @Override
    public Observable<JsonDocument> call(String id) {
        return bucket.async().get(id);
    }
})
.toList()
.toBlocking()
.single();

  2. Create a design view, then retrieve a subset of its index with startKey and endKey

    // Map function
function(doc, meta) {
    if (doc.type == 'otherdoc') {
        emit(meta.id, doc);
    }
}

// ViewQuery (in a java method)
ViewQuery.from('designOther', 'viewOther')
  .startKey('otherdoc1')
  .endKey('otherdoc2');

Upvotes: 2

Views: 1784

Answers (1)

Simon Basl&#233;
Simon Basl&#233;

Reputation: 28301

in Couchbase, when you know the key, the SDK knows which node to ask for that key (through hashing). On the other hand, querying a view implies the view engine to contact every node in the cluster.

So direct get && batching in RxJava, since you know the key(s), will save you additional round-trips and should end up the better performer!

Upvotes: 2

Related Questions