批量插入千万数据到Elasticsearch之bulkProcessor_bulkprocessor使用(2)

        @Override
        public void afterBulk(long executionId, BulkRequest request, BulkResponse response) {
            // 执行之后调用
            System.out.println("成功" + request.numberOfActions() + "条数据，用时" + response.getTook());
        }

        @Override
        public void afterBulk(long executionId, BulkRequest request, Throwable failure) {
            // 失败时调用
            System.out.println("失败" + request.numberOfActions() + "条数据");
            System.out.println("失败" + failure);
        }
    };
    BulkProcessor bulkProcessor = BulkProcessor.builder(
            (request, bulkListener) -> client.bulkAsync(request, RequestOptions.DEFAULT, bulkListener),
            listener)
            .setBulkActions(5000)
            .setBulkSize(new ByteSizeValue(5L, ByteSizeUnit.MB))
            .setFlushInterval(TimeValue.timeValueSeconds(10L))
            .setConcurrentRequests(10)
            .setBackoffPolicy(BackoffPolicy.exponentialBackoff(TimeValue.timeValueMillis(100), 3))
            .build();

    SearchRequest searchRequest = new SearchRequest(index);
    SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
    searchSourceBuilder.query(QueryBuilders.termsQuery("userNo", userNos));
    searchSourceBuilder.size(10000);
    searchRequest.source(searchSourceBuilder);

    SearchResponse searchResponse = client.search(searchRequest, RequestOptions.DEFAULT);
    if (searchResponse.getHits().getTotalHits().value == 0) {
        return;
    }
    if (searchResponse.getHits().getTotalHits().value > 0) {
        for (SearchHit hit : searchResponse.getHits().getHits()) {
            bulkProcessor.add(new DeleteRequest(index).id(hit.getId()));
        }
        bulkProcessor.flush();
        bulkProcessor.close();
        client.close();
    }

2.遇到的问题连接超时和io问题，这是因为发起的线程太多了上面我设置了10个线程，一个线程5000的数据，.setBulkActions(5000)  .setConcurrentRequests(10) 因为线程太多导致有一个线程不成功的话就会导致缺少数据，后面解决把.setBulkActions(2000)  .setConcurrentRequests(1）就成功导入1000万数据成功了


![](https://img-blog.csdnimg.cn/078a7929b2534dad964bafb5f4906df3.png)


3.遇到的坑


最后一次如果没有达到5000的不会发送请求，手动刷新一次.flush,关闭客户端不能立马关闭不然会出现I/O异常，使用bulkProcessor.awaitClose( timeout: executeTime/1000,TimeUnit.SECoNDS);关闭，加在flush的后面一行



![888](https://img-blog.csdnimg.cn/55ce92a5861e40a1a38563af902ad01c.png)


 4.最终代码成功导入一千万数据到es完整代码耗时30分钟

public static void bulksaveByUserNoRequest(String index, List userNos) throws IOException {
//创建ES客户端
try (RestHighLevelClient client = getClient()) {
BulkProcessor.Listener listener = new BulkProcessor.Listener() {
@Override
public void beforeBulk(long executionId, BulkRequest request) {
// 执行之前调用
System.out.println(“操作” + request.numberOfActions() + “条数据”);
}

            @Override
            public void afterBulk(long executionId, BulkRequest request, BulkResponse response) {
                // 执行之后调用
                System.out.println("成功" + request.numberOfActions() + "条数据，用时" + response.getTook());
            }

            @Override
            public void afterBulk(long executionId, BulkRequest request, Throwable failure) {
                // 失败时调用
                System.out.println("失败" + request.numberOfActions() + "条数据");

既有适合小白学习的零基础资料，也有适合3年以上经验的小伙伴深入学习提升的进阶课程，涵盖了95%以上大数据知识点，真正体系化！

由于文件比较多，这里只是将部分目录截图出来，全套包含大厂面经、学习笔记、源码讲义、实战项目、大纲路线、讲解视频，并且后续会持续更新

需要这份系统化资料的朋友，可以戳这里获取

战项目、大纲路线、讲解视频，并且后续会持续更新**

需要这份系统化资料的朋友，可以戳这里获取