Performance experiment for apache/accumulo#275 and ACCUMULO-4667

LocGroupPerfTest.java

import java.io.BufferedOutputStream;
import java.io.File;
import java.io.FileOutputStream;
import java.util.LinkedHashSet;
import java.util.Map.Entry;
import java.util.Set;

import org.apache.accumulo.core.client.Scanner;
import org.apache.accumulo.core.client.rfile.RFile;
import org.apache.accumulo.core.client.rfile.RFileWriter;
import org.apache.accumulo.core.data.Key;
import org.apache.accumulo.core.data.Value;
import org.apache.hadoop.io.Text;

public class LocGroupPerfTest {
  public static void main(String[] args) throws Exception {
    String filename = createFile(100, 10, 10000, 20);
    scan(filename, 5, 340);
  }

  private static String toRow(int r) {
    return String.format("r%08d", r);
  }

  private static String toFamily(int f) {
    return String.format("f%04d", f);
  }

  private static String toQual(int q) {
    return String.format("q%04d", q);
  }


  private static void scan(String filename, int ... fams) throws Exception {
    Scanner scanner = RFile.newScanner().from(filename).build();

    for (int f : fams) {
      scanner.fetchColumnFamily(new Text(toFamily(f)));
    }


    for(int i = 0; i < 100; i++) {
      int count = 0;
      long t1 = System.currentTimeMillis();
      for (Entry<Key,Value> entry : scanner) {
        count++;
      }
      long t2 = System.currentTimeMillis();
      System.out.println(count+" "+(t2 -t1));
    }
  }

  private static String createFile(int rows, int groups, int families, int quals) throws Exception {

    String filename = String.format("test-r%d-g%d-f%d-q%d.rf", rows, groups, families, quals);
    if(new File(filename).exists()) {
      return filename;
    }

    BufferedOutputStream out = new BufferedOutputStream(new FileOutputStream(filename));
    RFileWriter writer = RFile.newWriter().to(out).build();

    int famsPerGroup = families / groups;

    for(int g = 0; g < groups; g++) {

      Set<String> familySet = new LinkedHashSet<>();
      if(g == groups - 1) {
        for(int f = g * famsPerGroup; f < families; f++) {
          familySet.add(toFamily(f));
        }

        writer.startDefaultLocalityGroup();
      } else {

        for(int f = g * famsPerGroup; f < (g+1)*famsPerGroup; f++) {
          familySet.add(toFamily(f));
        }

        writer.startNewLocalityGroup("lg-"+g, familySet);
      }

      //System.out.println("group : " + g);

      int count = 0;
      for (int r = 0; r < rows; r++) {
        for (String family : familySet) {
          for (int q = 0; q < quals; q++) {
            Key k = new Key(toRow(r), family, toQual(q), "", 42);
            writer.append(k, k.hashCode() + "");
            count++;
            // System.out.printf(" %s %s %s\n", toRow(r), family, toQual(q));
          }
        }
      }
      System.out.println("added "+count+" keys");
    }

    writer.close();


    return filename;
  }
}