/*
* 测试普通模式与PipeLine模式的效率：
* 测试方法：向redis中插入10000组数据
*/
public static void testPipeLineAndNormal(Jedis jedis)
throws InterruptedException {
Logger logger = Logger.getLogger("javasoft");
long start = System.currentTimeMillis();
for (int i = 0; i < 10000; i++) {
jedis.set(String.valueOf(i), String.valueOf(i));
}
long end = System.currentTimeMillis();
logger.info("the jedis total time is:" + (end - start));
Pipeline pipe = jedis.pipelined();// 先创建一个pipeline的链接对象
long start_pipe = System.currentTimeMillis();
for (int i = 0; i < 10000; i++) {
pipe.set(String.valueOf(i), String.valueOf(i));
}
pipe.sync();// 获取所有的response
long end_pipe = System.currentTimeMillis();
logger.info("the pipe total time is:" + (end_pipe - start_pipe));
BlockingQueue<String> logQueue = new LinkedBlockingQueue<String>();
long begin = System.currentTimeMillis();
for (int i = 0; i < 10000; i++) {
logQueue.put("i=" + i);
}
long stop = System.currentTimeMillis();
logger.info("the BlockingQueue total time is:" + (stop - begin));
}
从上述代码以及结果中可以明显的看到PipeLine在“批量处理”时的优势。
