Hibernate Search 5.8.0.cr1 发布了,Hibernate Search 的作用是对数据库中的数据进行检索的。它是 hibernate 对著名的全文检索系统 Lucene 的一个集成方案,作用在于对数据表中某些内容庞大的字段(如声明为text的字段)建立全文索引,这样通过
hibernate search 就可以对这些字段进行全文检索后获得相应的 POJO,从而加快了对内容庞大字段进行模糊搜索的速度( sql
语句中 like 匹配)。 改版本包含了在5.8.0.beta4中的 bug 修复和改进,这将是正式版发布前社区的测试和bug报告的最后机会。 该版本最主要的更新包括:
HSEARCH-2831: request signing for Amazon’s proprietary IAM authentication mechanism now requires you to set the hibernate.search.default.elasticsearch.aws.signing.enabled property to true , allowing you to easily disable signing even if the hibernate-search-elasticsearch-aws JAR is in your classpath. HSEARCH-2818 / HSEARCH-2821: sending requests to Elasticsearch is now much less memory-consuming. HSEARCH-2764: we improved the orchestration of index updates before they are sent to the Elasticsearch client: Index
updates originating from a single Hibernate Search node will now be
sent to Elasticsearch in the order they were generated, even when they
come from different threads. Mass indexing will now add
documents in parallel, allowing you to take advantage of having multiple
connections to the Elasticsearch cluster. Note you can customize the maximum number of connections using the hibernate.search.default.elasticsearch.max_total_connection and hibernate.search.default.elasticsearch.max_total_connection_per_route configuration properties. The internal index update queues are now bounded, thus performing mass indexing on very large data sets will not trigger an OutOfMemoryError anymore.
…… 完整内容请查看发行说明。 |