Elasticsearch Java API Batch processing

Original address: http://blog.csdn.net/u012116196/article/details/51754845

The Bulkprocessor class provides a simple interface for automatic flushing of bulk operations based on the number or size of requests, or at a given period.

Using it, you first create a Bulkprocessor instance:[Java]  View Plain  copy import org.elasticsearch.action.bulk.backoffpolicy;   import  org.elasticsearch.action.bulk.bulkprocessor;   Import org.elasticsearch.common.unit.bytesizeunit ;   import org.elasticsearch.common.unit.bytesizevalue;   import  org.elasticsearch.common.unit.timevalue;      bulkprocessor bulkprocessor =  Bulkprocessor.builder (           client,  //<1 >           new bulkprocessor.listener ()  {                 @Override                 public void beforebulk (long  executionid,                      &nbsP;                bulkrequest  request)  { ... } //<2>                    @Override                 public void afterbulk (long executionid,                                       bulkrequest request,                                        Bulkresponse response)  { ... } //<3>                    @Override                 public void afterbulk (long executionid,                                       BulkRequest request,                                      throwable  failure)  { ... } //<4>           &NBSP})            .setbulkactions (10000)  //<5>            .setbulksize (New bytesizevalue (1,  BYTESIZEUNIT.GB))  //<6>           .setflushinterval (TimeValue.timeValueSeconds ( 5)  //<7>           .setconcurrentrequests (1)  //<8>           .setbackoffpolicy (               backoffpolicy.exponentialbackoff ( Timevalue.timevaluemillis,  3)  //<9>            .build ();  

	Add your Elasticsearch customer
	Executed before this method is called. For example you can see Numberofactions request.numberofactions ()
	Call this method after bulk execution. For example, you can check if there are some failed requests