休眠搜索空结果与工作查询

Question

休眠：5.2.8。最终
休眠搜索：5.7.0。最终
组态：

spring.jpa.properties.hibernate.search.analyzer=org.apache.lucene.analysis.en.EnglishAnalyzer
logging.level.org.hibernate.search.fulltext_query=debug

我的代码：

FullTextEntityManager ftsEntityManager = Search.getFullTextEntityManager(entityManager);
        QueryBuilder queryBuilder = ftsEntityManager.getSearchFactory()
                .buildQueryBuilder().forEntity(Report.class).get();
        BooleanQuery.Builder builder = new BooleanQuery.Builder();
        BooleanQuery.Builder cmpBuilder = new BooleanQuery.Builder();
        companies.forEach(cmp -> cmpBuilder.add(
                NumericRangeQuery.newLongRange("merchant.company.company_id", cmp.getId(), cmp.getId(), true, true),
                BooleanClause.Occur.SHOULD));
        builder.add(cmpBuilder.build(), BooleanClause.Occur.MUST);
        builder.add(new TermQuery(new Term("status", status.name())), BooleanClause.Occur.MUST);
        org.apache.lucene.search.Query fromQuery = queryBuilder.range()
                .onField("createdAt")
                .above(from)
                .createQuery();
        org.apache.lucene.search.Query toQuery = queryBuilder.range()
                .onField("createdAt")
                .below(to)
                .createQuery();
        builder.add(fromQuery, BooleanClause.Occur.MUST);
        builder.add(toQuery, BooleanClause.Occur.MUST);           
        FullTextQuery fullTextQuery = ftsEntityManager.createFullTextQuery(builder.build(), Report.class);
        int results = fullTextQuery.getResultSize();
        EntityGraph graph = this.entityManager.getEntityGraph("Report.fetchMerchantAndAccount");
        List resultList = fullTextQuery
                .setFirstResult(pageable.getOffset())
                .setMaxResults(pageable.getPageSize())
                .setHint("javax.persistence.fetchgraph", graph)
                .getResultList();

记录查询：

2017-03-24 02:39:50.942 DEBUG 1984 --- [tp1560534902-96] org.hibernate.search.fulltext_query      : HSEARCH000274: Executing Lucene query '+(merchant.company.company_id:[2 TO 2]) +status:COMPLETE +createdAt:[0000020161102000000000000000 TO *] +createdAt:[* TO 0000020170324235959999999999]'

我已经在Luke中测试了此查询，并且效果很好： 看起来像这个参数中的问题+(merchant.company.company_id:[2 TO 2]) +status:COMPLETE但是我不知道出了什么问题。

Answer 1

问题是我应该使用QueryBuilder而不是Lucene类。 因此，更改此：

companies.forEach(cmp -> cmpBuilder.add(
                NumericRangeQuery.newLongRange("merchant.company.company_id", cmp.getId(), cmp.getId(), true, true),
                BooleanClause.Occur.SHOULD));

对此：

companies.forEach(cmp -> cmpBuilder.add(queryBuilder.range().onField("merchant.company.company_id")
                .from(cmp.getId()).to(cmp.getId()).createQuery(), BooleanClause.Occur.SHOULD));

和这个：

builder.add(new TermQuery(new Term("status", status.name())), BooleanClause.Occur.MUST);

对此：

builder.add(queryBuilder.keyword().onField("status").matching(status).createQuery(), BooleanClause.Occur.MUST);

解决了一个问题！