简体繁体 English

使用多个单词的Apache Solr通配符搜索

[英]Apache solr wild card searching with multiple words

原文 2016-07-01 10:29:02 5 1 solr

We are using apache solr with php. 我们正在使用apache solr和php。

There is a problem in wild card searching. 通配符搜索中存在问题。

We want to search "project manage*" which can list possible results like project manager, project management etc. However, it is not working whenever there are two words in wild card searching 我们想搜索“项目管理*”，它可以列出可能的结果，例如项目经理，项目管理等。但是，只要通配符搜索中有两个单词，它就无法工作

For example "projectmanage*" is working whereas "proejct manage*" is not working. 例如，“ projectmanage *”在工作，而“ proejct manage *”在行。 We also tried by escaping the space but it is not working either.. 我们也尝试过逃避空间，但是它也不起作用。

Looking forward to all valuable inputs.. thanks in advance. 期待所有宝贵的意见。

1 个解决方案

When applying a wild card, the regular analysis chain is not performed when querying. 应用通配符时，查询时不执行常规分析链。 This results in Solr looking for tokens starting with with "project manage" - and if you have an analysis chain when indexing, your text is usually split into multiple tokens. 这样一来，Solr就会以“项目管理”开始查找令牌-如果在建立索引时具有分析链，则通常会将文本拆分成多个令牌。

You can use a Shingle filter to index multiple tokens as a single token, which can be used to get around the issue (be sure to use the same separator as you use in your text). 您可以使用Shingle筛选器将多个标记作为单个标记编制索引，该标记可用于解决问题（请确保使用与文本中相同的分隔符）。

Another option is to lowercase the field when indexing and querying and use a regular StrField which isn't processed in any way, or use a KeywordTokenizer - which keeps the indexed content as a single token. 另一种选择是在索引和查询时小写该字段，并使用未经任何方式处理的常规StrField，或使用KeywordTokenizer-将索引内容保留为单个标记。