JAVA使用Elasticsearch7.x查询的BoolQueryBuilder写法

Elasticsearch7.x集成网上一大堆,随便百度就行

此文章仅针对写分页查询时传入各种参数,然后的不同写法,仅供参考

实体中的部分代码

@Document(indexName = "appeal_doc")
@Data
@AllArgsConstructor
@NoArgsConstructor
@Builder
public class AppealDoc implements Serializable {


    @Id
    @Field(type = FieldType.Text)
    private String id;

    @Field(type = FieldType.Keyword)
    private String city;

    ......
    ......
    ......

    
    @Field(type = FieldType.Date)
    @JsonFormat(pattern = "yyyy-MM-dd HH:mm:ss", timezone = "GMT+8") 
    private Date appealDate;

    
    @Field(type = FieldType.Text)
    private String appealContent;

}

分页查询的部分代码

service上注释也写了,仅针对实体上Field类型为Text,Keyword,Date三种作为示例

BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery();
        // keyword不分词,相当于: t.city  = 'cities'
        if(StrUtil.isNotEmpty(query.getCities())) {
            String cities = query.getCities();
            boolQueryBuilder.must(QueryBuilders.termQuery("city", cities));
        }
        // keyword不分词,相当于 and t.complaintEnterprise in ( 'enterprises[0]', 'enterprises[1]', ...)
        if(StrUtil.isNotEmpty(query.getEnterprises())) {
            String[] enterprises = query.getEnterprises().split(",");
            boolQueryBuilder.must(QueryBuilders.termsQuery("complaintEnterprise", enterprises));
        }
        // 相当于是 like "categoryCode%"
        if(StrUtil.isNotEmpty(query.getCategoryCode())) {
            String categoryCode = query.getCategoryCode();
            boolQueryBuilder.must(QueryBuilders.wildcardQuery("categoryCodeNames", categoryCode+"*"));
        }

        // 相当于是 like "username%"
        if(StrUtil.isNotEmpty(query.getUsername())) {
            String username = query.getUsername();
            boolQueryBuilder.must(QueryBuilders.wildcardQuery("username", username+"*"));
        }

        // 相当于是 like "businessCode%"
        if(StrUtil.isNotEmpty(query.getBusinessCode())) {
            String businessCode = query.getBusinessCode();
            boolQueryBuilder.must(QueryBuilders.wildcardQuery("businessCodeNames", businessCode+"*"));
        }

        // Text自动分词,matchPhrase:模糊匹配且完整
        // 相当于:  and ( t.appealContent = '%searchKey[0]%' or t.appealContent = '%searchKey[1]%' or...)
        String searchKey = query.getHotWords();
        if(StrUtil.isNotEmpty(searchKey)) {
//            boolQueryBuilder.must(QueryBuilders.matchPhraseQuery("appealContent", searchKey));
            BoolQueryBuilder boolQueryBuilder2 = QueryBuilders.boolQuery();
            for (String s : searchKey.split(",")) {
                boolQueryBuilder2.should(QueryBuilders.matchPhraseQuery("appealContent", s));
            }
            boolQueryBuilder.must(boolQueryBuilder2);
        }

        String appealContent = query.getAppealContent();
        if(StrUtil.isNotEmpty(appealContent)) {
            boolQueryBuilder.must(QueryBuilders.matchPhraseQuery("appealContent", appealContent));
        }

        // 申诉时间 相当于:  and t.appealDate between #{startTime} and #{endTime}
        if (query.getAppealStartDate() != null && query.getAppealEndDate() != null ){
            Date startTime = query.getAppealStartDate();
            Date endTime = query.getAppealEndDate();
            boolQueryBuilder.must(QueryBuilders.rangeQuery("appealDate").from(startTime).to(endTime).includeUpper(true).includeLower(true));
        }

        // ES分页搜索
        NativeSearchQuery nativeSearchQuery = new NativeSearchQueryBuilder()
                //查询条件
                .withQuery(boolQueryBuilder)
                //分页,page从0开始
                .withPageable(PageRequest.of(pageNum, pageSize))
                //排序
                .withSorts(SortBuilders.fieldSort("appealDate").order(SortOrder.DESC))
                //高亮字段显示
                .withHighlightFields(
                        new HighlightBuilder.Field("appealContent").preTags("<em>").postTags("</em>")
                )
                .build();
        SearchHits<AppealDoc> docs = elasticsearchRestTemplate.search(nativeSearchQuery, AppealDoc.class);

部分使用后的感觉

1. must  => and

2. should => or 

3. 高亮显示的仅是当前字段的某部分数据(存在搜索条件的前后语句而已),并非全部

4.时间如果使用LocalDate/LocatDateTime反正是各种坑,自己转一次Date感觉方便点

5.分页查询时pageNum/curent/pages(当前页)是以0开始的

6.保存和更新是一起的