elasticsearch - Elasticsearch多个分析器不起作用

我有一个标题如下的小测试规:
' Santos(SAN)XAvaíFC(AVAI)-Canindé(圣保罗)'

如果我尝试搜索“ san ”或“ avai ”或“avaí”或“ santos ”之类的词，则效果很好。但是，如果我尝试搜索“ Santos(SAN)XAvaíFC(AVAI)”，则它应该仅返回1 reg，并且返回所有3 reg。

这是测试数据:
https://gist.github.com/PtkFerraro/83c4b693cf770c3320fe0530a4e1ddc7

这是分析仪和映射
https://gist.github.com/PtkFerraro/eb3244bf8c589b234a13d7f2b693cf77

搜索如下:
https://gist.github.com/PtkFerraro/c0f8ed300566cce3b5118fff1522a421

提前致谢

{
    "settings": {
        "analysis": {
            "analyzer": {
                "title_default_analyzer": {
                    "type": "custom",
                    "tokenizer": "title_tokenizer",
                    "filter": ["lowercase","brazilian_filter","asciifolding"
                    ]
                },
                "title_snowball_analyzer": {
                    "type": "custom",
                    "tokenizer": "title_tokenizer",
                    "filter": ["lowercase","brazilian_filter","asciifolding","snowball"
                    ]
                },
                "title_shingle_analyzer": {
                    "type": "custom",
                    "tokenizer": "title_tokenizer",
                    "filter": ["lowercase","brazilian_filter","shingle","asciifolding"
                    ]
                },
                "title_ngram_analyzer": {
                    "type": "custom",
                    "tokenizer": "title_tokenizer",
                    "filter": ["lowercase","brazilian_filter","asciifolding","edge_ngram_filter"
                    ]
                },
                "title_search_analyzer": {
                    "type": "custom",
                    "tokenizer": "title_tokenizer",
                    "filter": ["lowercase","brazilian_filter","asciifolding"
                    ]
                }
            },
            "filter": {
                "brazilian_filter": {
                    "type": "stemmer",
                    "name": "brazilian",
                    "token_chars": ["letter", "digit"]
                },
                "edge_ngram_filter": {
                    "type": "edgeNGram",
                    "min_gram": 3,
                    "max_gram": 50,
                    "token_chars": ["letter", "digit"]
                }
            },
            "tokenizer": {
                "title_tokenizer": {
                    "type": "letter"
                }
            }
        }
    },
    "mappings": {
        "entersport": {
            "_all": {
                "enabled": false
            },
            "properties": {
                "is_adult": {
                    "type": "boolean"
                },
                "match_start": {
                    "type": "date"
                },
                "match_title": {
                    "type": "text",
                    "fields": {
                        "title": {
                            "type": "text",
                            "analyzer": "title_default_analyzer"
                        },
                        "snowball": {
                            "type": "text",
                            "analyzer": "title_snowball_analyzer"
                        },
                        "shingles": {
                            "type": "text",
                            "analyzer": "title_shingle_analyzer"
                        },
                        "ngrams": {
                            "type": "text",
                            "analyzer": "title_ngram_analyzer",
                            "search_analyzer": "title_search_analyzer"
                        }
                    }
                }
            }
        }
    }
}

最佳答案

看来您想支持两种不同的搜索:(1)将san或avai这样的单个词匹配到足球比赛标题，以及(2)精确词组匹配部分或全部比赛标题。

您正在使用的query string query被解析为单独的术语，因此在您的示例中Santos (SAN) X Avaí FC (AVAI)包含X token ，该 token 与所有3个示例文档匹配，因为它们每个都包含X。

您的查询:

{
    "query_string": {
        "query": "Santos (SAN) X Avaí FC (AVAI)",
        "fields": [
            "title^10",
            "match_title.snowball^2",
            "match_title.shingles^2",
            "match_title.ngrams"
        ]
    }
}

如果要匹配整个短语，则需要使用查询来完成。评论中建议使用Match phrase queries。您还可以使用匹配查询，将operator设置为and查询中的术语。这将允许您使用相同的查询类型来匹配avai和Santos (SAN) X Avaí FC (AVAI)。

我认为您可能更喜欢以下内容:

{
  "query": {
    "multi_match" : {
      "query":      "Santos (SAN) X Avaí FC (AVAI)",
      "fields":     [
            "title^10",
            "match_title.snowball^2",
            "match_title.shingles^2",
            "match_title.ngrams"
        ],
        "operator":   "and"
    }
  }
}

关于elasticsearch - Elasticsearch多个分析器不起作用，我们在Stack Overflow上找到一个类似的问题：https://stackoverflow.com/questions/43902570/