redis pipelineを使用してhashデータを一括挿入

import org.apache.commons.lang.math.RandomUtils;

import redis.clients.jedis.Jedis;
import redis.clients.jedis.Pipeline;

public class RedisEasyTest {

	private static Jedis jedis = new Jedis("xx.xx.xx.xx");

	private static Pipeline p = jedis.pipelined();

	private static int KEY_COUNT = 10000;

	private static int FIELD_COUNT = 10;

	public void single() {
		for (int i = 0; i < KEY_COUNT; i++) {
			String key = RandomUtils.nextInt(5) + "";
			for (int j = 0; j < FIELD_COUNT; j++) {
				jedis.hset(key, j + "", i + j + "");
				jedis.expire(key, 3600);
			}
		}
	}

	public void batch() {
		int index = 0;
		for (int i = 0; i < KEY_COUNT; i++) {
			String key = RandomUtils.nextInt(5) + "";
			for (int j = 0; j < FIELD_COUNT; j++) {
				p.hset(key, j + "", i + j + "");
				p.expire(key, 3600);
			}
			if (++index % 1000 == 0) {
				p.sync();
			}
		}
		p.sync();
	}

	public static void main(String[] args) {
		long start = System.currentTimeMillis();
		RedisEasyTest r = new RedisEasyTest();
		r.single();
		System.out.printf("single use %d sec 
", (System.currentTimeMillis() - start) / 1000);
		start = System.currentTimeMillis();
		r.batch();
		System.out.printf("batch use %d sec 
", (System.currentTimeMillis() - start) / 1000);

	}
}

出力結果:
single use 30 sec batch use 0 sec
 
pipelineによる一括挿入データのパフォーマンスは非常に良好であることがわかります.
 
 
--end