回答你问题的最后一部分:solr 还有一个 ngram 过滤器.所以你不应该使用 ngram 分词器(例如像WhitespaceTokenizer"这样的分词器)，应用所有 pre-ngram 过滤器，然后添加这个:

To answer to the last part of your question: solr has also an ngram filter. So you should not use the ngram tokenizer (but one like "WhitespaceTokenizer" for example), apply all pre-ngram filters and then add this one:

<filter class="solr.NGramFilterFactory" minGramSize="2" maxGramSize="3" />