import org.apache.http.HttpHost;
import org.elasticsearch.action.search.SearchRequest;
import org.elasticsearch.action.search.SearchResponse;
import org.elasticsearch.client.RequestOptions;
import org.elasticsearch.client.RestHighLevelClient;
import org.elasticsearch.index.query.QueryBuilders;
import org.elasticsearch.search.builder.SearchSourceBuilder;
import org.elasticsearch.search.sort.SortOrder;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
public class ElasticsearchSearchAfterExample {
private final RestHighLevelClient client;public ElasticsearchSearchAfterExample(RestHighLevelClient client) {this.client = client;
}/*** 使用 search_after 实现分页查询，并每次都重新请求总文档数。** @param indexName 索引名称* @param pageSize 每页大小* @param sortField 排序字段* @param pageNum 页码 (从1开始)* @param queryMap 查询条件映射* @return 包含当前页文档列表和总文档数的 Map*/
public Map<String, Object> paginateWithSearchAfter(String indexName, int pageSize, String sortField, int pageNum, Map<String, Object> queryMap) throws IOException {List<Map<String, Object>> pageDocs = new ArrayList<>();Long totalHits = null;// 构建查询对象SearchSourceBuilder sourceBuilder = new SearchSourceBuilder();sourceBuilder.query(buildQueryFromMap(queryMap));sourceBuilder.size(pageSize);sourceBuilder.sort(sortField, SortOrder.ASC);// 先获取总文档数SearchRequest countRequest = new SearchRequest(indexName);SearchSourceBuilder countSourceBuilder = new SearchSourceBuilder();countSourceBuilder.query(buildQueryFromMap(queryMap));countSourceBuilder.size(0); // 只关心总数，不需要返回具体文档countSourceBuilder.trackTotalHits(true); // 确保跟踪总命中数countRequest.source(countSourceBuilder);SearchResponse countResponse = client.search(countRequest, RequestOptions.DEFAULT);totalHits = countResponse.getHits().getTotalHits().value;// 构建分页查询请求SearchRequest searchRequest = new SearchRequest(indexName);searchRequest.source(sourceBuilder);// 如果不是第一页，则需要设置 search_after 参数if (pageNum > 1) {// 获取上一页最后一个文档的排序值，用于构建 search_after 请求// 注意：这里简化了处理，实际应用中需要根据具体情况实现// 这里假设有一个方法 getLastSortValues 来获取这些值List<Object> lastSortValues = getLastSortValues(pageNum - 1, indexName, pageSize, sortField, queryMap);if (lastSortValues != null && !lastSortValues.isEmpty()) {sourceBuilder.searchAfter(lastSortValues.toArray(new Object[0]));}}SearchResponse searchResponse = client.search(searchRequest, RequestOptions.DEFAULT);for (SearchHit hit : searchResponse.getHits()) {pageDocs.add(hit.getSourceAsMap());}// 返回包含文档列表和总文档数的结果return Map.of("docs", pageDocs, "totalHits", totalHits);
}/*** 根据查询条件映射构建查询对象。*/
private org.elasticsearch.index.query.QueryBuilder buildQueryFromMap(Map<String, Object> queryMap) {// 示例：假设你有一个名为 "field" 的查询参数用于匹配特定字段if (queryMap == null || queryMap.isEmpty()) {return QueryBuilders.matchAllQuery();}return QueryBuilders.termQuery((String) queryMap.getOrDefault("field", ""), queryMap.getOrDefault("value", ""));
}/*** 获取指定页码前一页最后一个文档的排序值。*/
private List<Object> getLastSortValues(int pageNum, String indexName, int pageSize, String sortField, Map<String, Object> queryMap) throws IOException {// 这是一个简化的例子，实际应用中可能需要更复杂的逻辑来获取正确的排序值// 这里假设只需要取最后一条记录的排序值作为 search_after 的参数SearchRequest searchRequest = new SearchRequest(indexName);SearchSourceBuilder sourceBuilder = new SearchSourceBuilder();sourceBuilder.query(buildQueryFromMap(queryMap));sourceBuilder.size(1);sourceBuilder.sort(sortField, SortOrder.DESC); // 按降序排列以获取最后一条记录if (pageNum > 1) {// 计算实际需要跳过的文档数量并遍历到正确的起始位置...// （这部分代码与之前的例子相同）int skipDocs = (pageNum - 1) * pageSize;sourceBuilder.from(skipDocs);}searchRequest.source(sourceBuilder);SearchResponse searchResponse = client.search(searchRequest, RequestOptions.DEFAULT);if (!searchResponse.getHits().getHits().isEmpty()) {SearchHit lastHit = searchResponse.getHits().getHits()[0];return List.of(lastHit.getSortValues());}return null;
}public static void main(String[] args) throws IOException {try (RestHighLevelClient client = new RestHighLevelClient(RestClient.builder(new HttpHost("localhost", 9200, "http")))) {ElasticsearchSearchAfterExample example = new ElasticsearchSearchAfterExample(client);// 获取第1页的数据及总数，带查询参数Map<String, Object> queryMap = Map.of("field", "status", "value", "active");Map<String, Object> resultPage1 = example.paginateWithSearchAfter("my-index-000001", 10, "timestamp", 1, queryMap);System.out.println("Documents on Page 1: " + resultPage1.get("docs"));System.out.println("Total Hits: " + resultPage1.get("totalHits"));// 获取第2页的数据及总数，带相同的查询参数Map<String, Object> resultPage2 = example.paginateWithSearchAfter("my-index-000001", 10, "timestamp", 2, queryMap);System.out.println("Documents on Page 2: " + resultPage2.get("docs"));System.out.println("Total Hits: " + resultPage2.get("totalHits"));// 获取第1页的数据及总数，带不同查询参数Map<String, Object> queryMap2 = Map.of("field", "status", "value", "inactive");Map<String, Object> resultPage1Inactive = example.paginateWithSearchAfter("my-index-000001", 10, "timestamp", 1, queryMap2);System.out.println("Documents on Page 1 (Inactive): " + resultPage1Inactive.get("docs"));System.out.println("Total Hits (Inactive): " + resultPage1Inactive.get("totalHits"));}
}
}