Revision 57712
Added by Miriam Baglioni over 4 years ago
modules/dnet-mapreduce-jobs/trunk/src/main/java/eu/dnetlib/data/mapreduce/hbase/propagation/compositekeys/NaturalInstOrgKeyPartitioner.java | ||
---|---|---|
1 |
package eu.dnetlib.data.mapreduce.hbase.propagation.compositekeys; |
|
2 |
|
|
3 |
import org.apache.hadoop.io.Text; |
|
4 |
import org.apache.hadoop.mapreduce.Partitioner; |
|
5 |
|
|
6 |
/** |
|
7 |
* Created by miriam on 17/08/2018. |
|
8 |
*/ |
|
9 |
public class NaturalInstOrgKeyPartitioner extends Partitioner<InstOrgKey, Text > { |
|
10 |
|
|
11 |
@Override |
|
12 |
public int getPartition(InstOrgKey key, Text val, int numPartitions) { |
|
13 |
final int res = Math.abs(key.getId().hashCode() % numPartitions); |
|
14 |
|
|
15 |
return res; |
|
16 |
} |
|
17 |
} |
modules/dnet-mapreduce-jobs/trunk/src/main/java/eu/dnetlib/data/mapreduce/hbase/propagation/compositekeys/NaturalInstOrgKeyGroupingComparator.java | ||
---|---|---|
1 |
package eu.dnetlib.data.mapreduce.hbase.propagation.compositekeys; |
|
2 |
|
|
3 |
import org.apache.hadoop.io.WritableComparable; |
|
4 |
import org.apache.hadoop.io.WritableComparator; |
|
5 |
|
|
6 |
/** |
|
7 |
* Created by miriam on 17/08/2018. |
|
8 |
*/ |
|
9 |
public class NaturalInstOrgKeyGroupingComparator extends WritableComparator { |
|
10 |
|
|
11 |
protected NaturalInstOrgKeyGroupingComparator() { |
|
12 |
super(InstOrgKey.class, true); |
|
13 |
} |
|
14 |
|
|
15 |
@Override |
|
16 |
public int compare(WritableComparable w1, WritableComparable w2) { |
|
17 |
final InstOrgKey k1 = (InstOrgKey) w1; |
|
18 |
final InstOrgKey k2 = (InstOrgKey) w2; |
|
19 |
|
|
20 |
return k1.getId().compareTo(k2.getId()); |
|
21 |
} |
|
22 |
} |
modules/dnet-mapreduce-jobs/trunk/src/main/java/eu/dnetlib/data/mapreduce/hbase/propagation/compositekeys/InstOrgKey.java | ||
---|---|---|
1 |
package eu.dnetlib.data.mapreduce.hbase.propagation.compositekeys; |
|
2 |
|
|
3 |
import com.google.common.collect.ComparisonChain; |
|
4 |
import org.apache.hadoop.io.IntWritable; |
|
5 |
import org.apache.hadoop.io.Text; |
|
6 |
import org.apache.hadoop.io.WritableComparable; |
|
7 |
|
|
8 |
import java.io.DataInput; |
|
9 |
import java.io.DataOutput; |
|
10 |
import java.io.IOException; |
|
11 |
|
|
12 |
import static eu.dnetlib.data.mapreduce.hbase.propagation.PropagationConstants.*; |
|
13 |
|
|
14 |
/** |
|
15 |
* Created by miriam on 17/08/2018. |
|
16 |
*/ |
|
17 |
public class InstOrgKey implements WritableComparable<InstOrgKey> { |
|
18 |
|
|
19 |
|
|
20 |
|
|
21 |
private IntWritable keyType; |
|
22 |
|
|
23 |
private Text id; |
|
24 |
|
|
25 |
public InstOrgKey() {} |
|
26 |
|
|
27 |
public static InstOrgKey create(final int keyType, final String id) { |
|
28 |
return new InstOrgKey(keyType, id); |
|
29 |
} |
|
30 |
|
|
31 |
public static InstOrgKey datasource(final String id) { |
|
32 |
return new InstOrgKey(DATASOURCE, id); |
|
33 |
} |
|
34 |
|
|
35 |
public static InstOrgKey organization(final String id) { |
|
36 |
return new InstOrgKey(ORGANIZATION, id); |
|
37 |
} |
|
38 |
|
|
39 |
public static InstOrgKey publication(final String id){ |
|
40 |
return new InstOrgKey(PUBLICATION, id); |
|
41 |
} |
|
42 |
|
|
43 |
public InstOrgKey(final int keyType, final String id) { |
|
44 |
this.id = new Text(id); |
|
45 |
this.keyType = new IntWritable(keyType); |
|
46 |
} |
|
47 |
|
|
48 |
public void setKeyType(final IntWritable keyType) { |
|
49 |
this.keyType = keyType; |
|
50 |
} |
|
51 |
|
|
52 |
public void setId(final Text id) { |
|
53 |
this.id = id; |
|
54 |
} |
|
55 |
|
|
56 |
public Text getId() { |
|
57 |
return id; |
|
58 |
} |
|
59 |
|
|
60 |
public IntWritable getKeyType() { |
|
61 |
return keyType; |
|
62 |
} |
|
63 |
|
|
64 |
@Override |
|
65 |
public int compareTo(final InstOrgKey o) { |
|
66 |
final int res = ComparisonChain.start() |
|
67 |
.compare(getId(), o.getId()) |
|
68 |
.compare(getKeyType(), o.getKeyType()) |
|
69 |
.result(); |
|
70 |
|
|
71 |
//System.out.println(String.format("%s.compareTo(%s) = %s", toString(), o.toString(), res)); |
|
72 |
return res; |
|
73 |
} |
|
74 |
|
|
75 |
@Override |
|
76 |
public void write(final DataOutput out) throws IOException { |
|
77 |
keyType.write(out); |
|
78 |
id.write(out); |
|
79 |
} |
|
80 |
|
|
81 |
@Override |
|
82 |
public void readFields(final DataInput in) throws IOException { |
|
83 |
keyType = new IntWritable(); |
|
84 |
keyType.readFields(in); |
|
85 |
id = new Text(); |
|
86 |
id.readFields(in); |
|
87 |
} |
|
88 |
|
|
89 |
@Override |
|
90 |
public String toString() { |
|
91 |
return (new StringBuilder()) |
|
92 |
.append('{') |
|
93 |
.append(getKeyType().get()) |
|
94 |
.append(',') |
|
95 |
.append(getId()) |
|
96 |
.append('}') |
|
97 |
.toString(); |
|
98 |
} |
|
99 |
} |
Also available in: Unified diff
logic of the compositekeys for propagation moved on dedicated package. Ordering datasource, organization, results on the key that groups for the reducer