Skip to content

Commit bad60cc

Browse files
Merge pull request #19 from mohammadzainabbas/zain
Minor change
2 parents b79dd9b + fd7f0e5 commit bad60cc

File tree

2 files changed

+896
-894
lines changed

2 files changed

+896
-894
lines changed

src/main/java/exercise_4/Exercise_4.java

Lines changed: 8 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -44,29 +44,31 @@ public static void wikipedia(JavaSparkContext ctx, SQLContext sqlCtx) {
4444

4545
Utils.line_separator();
4646

47-
// For dumping factor
4847
List<Row> timeList = new ArrayList<Row>();
4948
StructType outputSchema = new StructType(new StructField[] {
50-
new StructField("dumping_factor", DataTypes.DoubleType, false, new MetadataBuilder().build()),
49+
new StructField("damping_factor", DataTypes.DoubleType, false, new MetadataBuilder().build()),
50+
new StructField("reset_probability (1 - damping_factor)", DataTypes.DoubleType, false, new MetadataBuilder().build()),
5151
new StructField("maxIter", DataTypes.IntegerType, false, new MetadataBuilder().build()),
5252
new StructField("time", DataTypes.LongType, false, new MetadataBuilder().build())
5353
});
54+
// For damping factor
5455
IntStream.range(1, 20).forEach(i -> {
5556
// For max iterations
5657
IntStream.range(1, 5).forEach(j -> {
57-
Double dumpFactor = i * 0.05;
58+
Double dampFactor = i * 0.05;
59+
Double resetProbability = 1 - dampFactor;
5860
Integer maxIteration = j * 5;
5961

6062
Long startTime = System.currentTimeMillis();
61-
GraphFrame gf = graphFrame.pageRank().resetProbability(dumpFactor).maxIter(maxIteration).run();
63+
GraphFrame gf = graphFrame.pageRank().resetProbability(resetProbability).maxIter(maxIteration).run();
6264
Long endTime = System.currentTimeMillis();
6365

6466
Long timeTaken = endTime - startTime;
65-
timeList.add(RowFactory.create(dumpFactor, maxIteration, timeTaken));
67+
timeList.add(RowFactory.create(dampFactor, resetProbability, maxIteration, timeTaken));
6668

6769
Dataset<Row> topVertices = gf.vertices().sort(org.apache.spark.sql.functions.desc("pagerank"));
6870

69-
String log = "dumping factor: '" + dumpFactor + "' maxIter: '" + maxIteration + "' time: '" + timeTaken + "' msec\n\n";
71+
String log = "damping factor: '" + dampFactor + "' maxIter: '" + maxIteration + "' time: '" + timeTaken + "' msec\n\n";
7072
Utils.print(log);
7173
topVertices.show(10);
7274
Utils.line_separator();

0 commit comments

Comments
 (0)