Elasticsearch-Search After分页-Java示例

Github源码：https://github.com/Mengzuozhu/es-demo

相关文章链接：

• from size 分页-Java示例
• Scroll 分页-Java示例
• Search After分页-Java示例

ES分页方式对比

参考：
https://blog.csdn.net/andybegin/article/details/83864171

5.6官方Search After文档

7.x官方Search After文档

分页方式	简述	缺点
from size	浅分页情况下，可以考虑使用；	深度分页情况下，性能较差
Scroll	基于历史快照和上一页的scroll_id，查询下一页；	● 滚屏翻页，快照比较占内存； ● 不适合实时搜索和跳页需求
Search After	1. 需要排序，官方（5.6版本）推荐使用 _uid 作为排序字段和全局唯一值（注：不同版本的ES，官方文档有变动，请参考对应版本的文档。比如，7.x版本建议拷贝_id字段值到另外一个支持doc value的字段，用于作为排序字段）； 2. 通过实时cursor来规避耗时和占存储的性能问题，基于上一页的结果帮助检索下一页； 3. 使用search_after参数时，from的值必须为0或者-1	● 需要排序； ● 不适合跳页需求

注：在7.x官方文档中，已不太建议使用Scroll API来深分页了https://www.elastic.co/guide/en/elasticsearch/reference/7.x/paginate-search-results.html#scroll-search-results

Java代码示例

ES 5.6.8版本

maven引用

		<dependency>
			<groupId>org.elasticsearch</groupId>
			<artifactId>elasticsearch</artifactId>
			<version>5.6.8</version>
		</dependency>
		<dependency>
			<groupId>org.elasticsearch.client</groupId>
			<artifactId>transport</artifactId>
			<version>5.6.8</version>
		</dependency>

Search After 的Java代码示例

import org.elasticsearch.action.search.SearchRequestBuilder;
import org.elasticsearch.action.search.SearchResponse;
import org.elasticsearch.client.transport.TransportClient;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.common.transport.InetSocketTransportAddress;
import org.elasticsearch.index.query.QueryBuilders;
import org.elasticsearch.search.SearchHit;
import org.elasticsearch.search.sort.SortOrder;
import org.elasticsearch.transport.client.PreBuiltTransportClient;

import java.net.InetAddress;
import java.net.UnknownHostException;
import java.util.function.Consumer;

public class EsSearchAfterHandler {

    private static final String UID = "_uid";
    
    public static void main(String[] args) throws UnknownHostException {
        // example: 127.0.0.1
        String esIp = "your es ip";
        TransportClient esClient = buildEsTransportClient("elasticsearch", esIp, 9301);
        int batchSize = 1000;
        SearchRequestBuilder searchRequestBuilder = esClient.prepareSearch("your index")
                .setQuery(QueryBuilders.matchAllQuery())
                .setSize(batchSize);
        EsSearchAfterHandler.searchForHit(searchRequestBuilder, searchHit -> {
            // your code ...
            // example:
            System.out.println(searchHit.getSourceAsString());
        });
    }

    
    /**
     * Search for hit.
     *
     * @param requestBuilder the request builder
     * @param consumer       the consumer
     */
    public static void searchForHit(SearchRequestBuilder requestBuilder, Consumer<SearchHit> consumer) {
        searchForResponse(requestBuilder, searchResponse -> {
            SearchHit[] searchHits = searchResponse.getHits().getHits();
            for (SearchHit searchHit : searchHits) {
                consumer.accept(searchHit);
            }
        });
    }

    /**
     * Search for response.
     *
     * @param requestBuilder the request builder
     * @param consumer       the consumer
     */
    public static void searchForResponse(SearchRequestBuilder requestBuilder, Consumer<SearchResponse> consumer) {
        if (requestBuilder == null || consumer == null) {
            return;
        }
        //requestBuilder.setSize(100); //在构建查询条件时，即可设置大小
        //设置排序字段
        SearchRequestBuilder sortBuilder = requestBuilder.addSort(UID, SortOrder.ASC);
        SearchResponse searchResponse = sortBuilder.get();
        SearchHit[] searchHits = searchResponse.getHits().getHits();
        while (searchHits.length > 0) {
            consumer.accept(searchResponse);
            SearchHit last = searchHits[searchHits.length - 1];
            sortBuilder = sortBuilder.searchAfter(last.getSortValues());
            searchResponse = sortBuilder.get();
            searchHits = searchResponse.getHits().getHits();
        }
    }
    
    private static TransportClient buildEsTransportClient(String esCluster, String esIp, Integer esPort)
            throws UnknownHostException {
        Settings esSettings = Settings.builder()
                .put("cluster.name", esCluster)
                .build();
        return new PreBuiltTransportClient(esSettings)
                .addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName(esIp), esPort));
    }
}

ES 7.9.2版本

maven引用

        <dependency>
            <groupId>org.elasticsearch</groupId>
            <artifactId>elasticsearch</artifactId>
            <version>7.9.2</version>
        </dependency>
        <dependency>
            <groupId>org.elasticsearch.client</groupId>
            <artifactId>elasticsearch-rest-high-level-client</artifactId>
            <version>7.9.2</version>
        </dependency>

Search After 的Java代码示例

/**
 * ES Search After查询
 *
 * @author zuozhu.meng
 **/
public class EsSearchAfterHandler {
    private static final String UID = "_id";

    /**
     * Search for hit.
     *
     * @param searchRequest the search request
     * @param consumer      the consumer
     */
    public static void searchForHit(RestHighLevelClient client, SearchRequest searchRequest,
                                    Consumer<SearchHit> consumer) throws IOException {
        searchForResponse(client, searchRequest, searchResponse -> forEachHits(searchResponse, consumer));
    }

    /**
     * Search for response.
     *
     * @param searchRequest the search request
     * @param consumer      the consumer
     */
    public static void searchForResponse(RestHighLevelClient client, SearchRequest searchRequest,
                                         Consumer<SearchResponse> consumer) throws IOException {
        if (searchRequest == null || consumer == null) {
            return;
        }
        SearchSourceBuilder sourceBuilder = searchRequest.source();
        // sourceBuilder.size(100); //在构建查询条件时，即可设置大小
        //设置排序字段
        sourceBuilder.sort(UID, SortOrder.ASC);
        SearchResponse searchResponse = client.search(searchRequest, RequestOptions.DEFAULT);
        SearchHit[] searchHits = searchResponse.getHits().getHits();
        while (searchHits.length > 0) {
            consumer.accept(searchResponse);
            SearchHit last = searchHits[searchHits.length - 1];
            sourceBuilder.searchAfter(last.getSortValues());
            searchResponse = client.search(searchRequest, RequestOptions.DEFAULT);
            searchHits = searchResponse.getHits().getHits();
        }
    }
    
    public static void forEachHits(SearchResponse searchResponse, Consumer<SearchHit> consumer) {
        if (searchResponse == null) {
            return;
        }
        SearchHit[] searchHits = searchResponse.getHits().getHits();
        for (SearchHit searchHit : searchHits) {
            consumer.accept(searchHit);
        }
    }
    
}