Which Analyzer to get results of appended words when data has space or dash in between?

问题

my mapping looks like as below. As you can see that Name field is analyzed

     {
       "state":"open",
       "settings":{
          "index":{
             "creation_date":"1453816191454",
             "number_of_shards":"5",
             "number_of_replicas":"1",
             "version":{
                "created":"1070199"
             },           
          }
       },
       "mappings":{
          "Product":{
             "properties":{
                "index":"not_analyzed",
                "store":true,
                "type":"string"
             },
            "Name":{                  
                "store":true,
                "type":"string"
             },            
             "Number":{
                "index":"not_analyzed",
                "store":true,
                "type":"string"
             },
             "id":{
                "index":"no",
                "store":true,
                "type":"integer"
   }
         }
      },
      "aliases":[

      ]
   }
}

When I query as below

   "query": {
            "match_phrase": {
               "Name": "hl-2240"
            }
         }

This works fine and also "hl 2240" works fine but when I type "hl2240". I dont get any results. I understand that it is because name is indexed as "hl-2240" and I guess that I am using standart or generic analyzer and it tokenizes as hl and 2240. while I dont have any token as hl2240 in reverse index, It doesnt find anything. I learnt that I should use another analyzer. But this is where I am stuck at. Which analyzer I can use? Should I reindex my index or I can use the Analyzer only to query? If I change the analyzer to index my data, I want to make sure that I am not loosing results for searching "hl-2240" or "hl 2240".

Update: Nest query I tried for Richa's answer.

   Client.CreateIndex("myIndex",
            ci => ci.Analysis(a => a.TokenFilters(f => f.Add("my_word_delimiter", new WordDelimiterTokenFilter
            {
              CatenateAll = true          

            }))
            .Analyzers(an => an.Add("my_analyzer", new CustomAnalyzer
            {
              Tokenizer = "whitespace",
              Filter = new List<string> {"standard",
              "lowercase",
              "my_word_delimiter"}
            }))));

回答1:

Try to use this analyzer:

{
 "settings": {
  "analysis": {
       "filter": {
        "my_word_delimiter": {
           "type": "word_delimiter",
           "catenate_all": true     <=== Notice this

        }
     },
     "analyzer": {
        "my_analyzer": {
           "type": "custom",
           "tokenizer": "whitespace",
           "filter": [
              "standard",
              "lowercase",
              "my_word_delimiter"
           ]
        }
     }

    }
  }
}

Read about catenate_all.

Use following command to see how string is being tokenized:

 curl -XGET "localhost:9200/index_8/_analyze?analyzer=my_analyzer&pretty=true" -d 'hl-2240'

THis will produce following output and hl-2240 will be indexed as

{
"tokens" : [ {
 "token" : "hl",
"start_offset" : 0,
"end_offset" : 2,
"type" : "word",
"position" : 0
}, {
"token" : "hl2240",
"start_offset" : 0,
"end_offset" : 7,
"type" : "word",
"position" : 0
}, {
"token" : "2240",
"start_offset" : 3,
"end_offset" : 7,
"type" : "word",
"position" : 1
 }  ]
}

Hope it helps you

来源：https://stackoverflow.com/questions/35697941/which-analyzer-to-get-results-of-appended-words-when-data-has-space-or-dash-in-b

标签

ElasticSearch

nest