Project

General

Profile

« Previous | Next » 

Revision 53604

implemented ConfigurableExportMapper

View differences:

modules/dnet-mapreduce-jobs/branches/master/src/main/java/eu/dnetlib/data/mapreduce/hbase/dataexport/ExportDuplicatesMapper.java
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
}
modules/dnet-mapreduce-jobs/branches/master/src/main/java/eu/dnetlib/data/mapreduce/hbase/dataexport/ConfigurableExportMapper.java
1
package eu.dnetlib.data.mapreduce.hbase.dataexport;
2

  
3
import com.google.common.base.Splitter;
4
import com.google.common.collect.Sets;
5
import com.googlecode.protobuf.format.JsonFormat;
6
import eu.dnetlib.data.mapreduce.util.DedupUtils;
7
import eu.dnetlib.data.proto.OafProtos;
8
import eu.dnetlib.data.proto.TypeProtos;
9
import org.apache.hadoop.hbase.client.Result;
10
import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
11
import org.apache.hadoop.hbase.mapreduce.TableMapper;
12
import org.apache.hadoop.io.Text;
13
import org.apache.hadoop.mapreduce.Mapper;
14

  
15
import java.io.IOException;
16
import java.util.Map;
17
import java.util.Set;
18

  
19
public class ConfigurableExportMapper extends TableMapper<Text, Text> {
20

  
21
    private static final String CF_DEFAULT = "result";
22
    private static final String QUALIFIER_DEFAULT = "body";
23

  
24
    private Text keyOut;
25
    private Text valueOut;
26

  
27
    private Set<String> columnFamilies = Sets.newHashSet();
28
    private Set<String> qualifiers = Sets.newHashSet();
29

  
30
    enum ExportMode {
31
        ALL,
32
        DUPS_ONLY
33
    }
34

  
35
    private ExportMode exportMode = ExportMode.ALL;
36

  
37
    @Override
38
    protected void setup(Context context) throws IOException, InterruptedException {
39
        keyOut = new Text("");
40
        valueOut = new Text();
41

  
42

  
43
        columnFamilies.addAll(Splitter.on(",").splitToList(context.getConfiguration().get("column.families", CF_DEFAULT)));
44
        qualifiers.addAll(Splitter.on(",").splitToList(context.getConfiguration().get("qualifiers", QUALIFIER_DEFAULT)));
45
        exportMode = ExportMode.valueOf(context.getConfiguration().get("export.mode", ExportMode.ALL.toString()));
46

  
47
    }
48

  
49
    @Override
50
    protected void map(ImmutableBytesWritable key, Result value, Context context) throws IOException, InterruptedException {
51
        for(String cf : columnFamilies) {
52
            switch (exportMode) {
53
                case ALL:
54
                    export(value, context, cf);
55
                    break;
56
                case DUPS_ONLY:
57
                    final Map<byte[], byte[]> mergedInMap = value.getFamilyMap(DedupUtils.getDedupCF_mergedInBytes(cf));
58
                    if (mergedInMap != null && !mergedInMap.isEmpty()) {
59
                        export(value, context, cf);
60
                    }
61

  
62
                    break;
63
            }
64
        }
65
    }
66

  
67
    private void export(Result value, Context context, String cf) throws IOException, InterruptedException {
68
        for(String q : qualifiers) {
69

  
70
            final byte[] data = value.getValue(cf.getBytes(), q.getBytes());
71

  
72
            if (data != null) {
73
                OafProtos.Oaf oaf = OafProtos.Oaf.parseFrom(data);
74
                valueOut.set(JsonFormat.printToString(oaf));
75
                context.write(keyOut, valueOut);
76

  
77
                context.getCounter(cf, q).increment(1);
78
            }
79
        }
80
    }
81

  
82
    
83
}

Also available in: Unified diff