forked from D-Net/dnet-hadoop
60 lines
1.5 KiB
Java
60 lines
1.5 KiB
Java
|
|
package eu.dnetlib.dhp.broker.oa.util.aggregators.withRels;
|
|
|
|
import org.apache.commons.lang3.StringUtils;
|
|
import org.apache.spark.sql.Encoder;
|
|
import org.apache.spark.sql.Encoders;
|
|
import org.apache.spark.sql.expressions.Aggregator;
|
|
|
|
import eu.dnetlib.broker.objects.OaBrokerMainEntity;
|
|
import scala.Tuple2;
|
|
|
|
public class AddDatasourceTypeAggregator
|
|
extends Aggregator<Tuple2<OaBrokerMainEntity, SimpleDatasourceInfo>, OaBrokerMainEntity, OaBrokerMainEntity> {
|
|
|
|
/**
|
|
*
|
|
*/
|
|
private static final long serialVersionUID = 8788588975496014728L;
|
|
|
|
@Override
|
|
public OaBrokerMainEntity zero() {
|
|
return new OaBrokerMainEntity();
|
|
}
|
|
|
|
@Override
|
|
public OaBrokerMainEntity finish(final OaBrokerMainEntity g) {
|
|
return g;
|
|
}
|
|
|
|
@Override
|
|
public OaBrokerMainEntity reduce(final OaBrokerMainEntity g,
|
|
final Tuple2<OaBrokerMainEntity, SimpleDatasourceInfo> t) {
|
|
final OaBrokerMainEntity res = StringUtils.isNotBlank(g.getOpenaireId()) ? g : t._1;
|
|
if (t._2 != null && StringUtils.isNotBlank(t._2.getType())) {
|
|
res.setCollectedFromType(t._2.getType());
|
|
}
|
|
return res;
|
|
|
|
}
|
|
|
|
@Override
|
|
public OaBrokerMainEntity merge(final OaBrokerMainEntity g1, final OaBrokerMainEntity g2) {
|
|
if (StringUtils.isNotBlank(g1.getOpenaireId()) && StringUtils.isNotBlank(g1.getCollectedFromType())) {
|
|
return g1;
|
|
} else {
|
|
return g2;
|
|
}
|
|
}
|
|
|
|
@Override
|
|
public Encoder<OaBrokerMainEntity> bufferEncoder() {
|
|
return Encoders.bean(OaBrokerMainEntity.class);
|
|
}
|
|
|
|
@Override
|
|
public Encoder<OaBrokerMainEntity> outputEncoder() {
|
|
return Encoders.bean(OaBrokerMainEntity.class);
|
|
}
|
|
}
|