Is it more advised to use rest client from elastichsearch-hadoop library for searching in elasticsearch instead of using the java client?
Or let me put it this way, what advantages does rest client from the es-hadoop library provide over the standard java client (for searching in elasticsearch not for ingesting data)?
By rest client I mean https://github.com/elastic/elasticsearch-hadoop/blob/master/mr/src/main/java/org/elasticsearch/hadoop/rest/RestClient.java
The question refers to the following versions: elastichsearch-hadoop 2.2.0-beta1 and the elasticsearch 2.0 java api.
The "rest client" in ES-Hadoop is not meant for public consumption. Hence why there are no docs for it and why it's not advertised anywhere.
Thus the answer to your question is "don't use it'. ES-Hadoop is meant to be used through the various libraries/runtimes it integrates with.
If that's not good enough/working for you, any other official client is a good alternative.