1
|
package eu.dnetlib.data.mapreduce.hbase.dataexport;
|
2
|
|
3
|
import com.googlecode.protobuf.format.JsonFormat;
|
4
|
import eu.dnetlib.data.mapreduce.util.DedupUtils;
|
5
|
import eu.dnetlib.data.proto.OafProtos;
|
6
|
import eu.dnetlib.data.proto.TypeProtos;
|
7
|
import org.apache.hadoop.hbase.client.Result;
|
8
|
import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
|
9
|
import org.apache.hadoop.hbase.mapreduce.TableMapper;
|
10
|
import org.apache.hadoop.io.Text;
|
11
|
|
12
|
import java.io.IOException;
|
13
|
import java.util.Map;
|
14
|
|
15
|
public class ExportDuplicatesMapper extends TableMapper<Text, Text> {
|
16
|
|
17
|
private Text keyOut;
|
18
|
private Text valueOut;
|
19
|
|
20
|
@Override
|
21
|
protected void setup(Context context) throws IOException, InterruptedException {
|
22
|
keyOut = new Text("");
|
23
|
valueOut = new Text();
|
24
|
}
|
25
|
|
26
|
@Override
|
27
|
protected void map(ImmutableBytesWritable key, Result value, Context context) throws IOException, InterruptedException {
|
28
|
|
29
|
final Map<byte[], byte[]> mergedInMap = value.getFamilyMap(DedupUtils.getDedupCF_mergedInBytes(TypeProtos.Type.result));
|
30
|
|
31
|
if (mergedInMap != null && !mergedInMap.isEmpty()) {
|
32
|
final byte[] body = value.getValue("result".getBytes(), DedupUtils.BODY_B);
|
33
|
|
34
|
if (body != null) {
|
35
|
OafProtos.Oaf oaf = OafProtos.Oaf.parseFrom(body);
|
36
|
valueOut.set(JsonFormat.printToString(oaf));
|
37
|
context.write(keyOut, valueOut);
|
38
|
}
|
39
|
}
|
40
|
|
41
|
}
|
42
|
}
|