Introduction to mapping in ES

Default mapping

elasticsearch( hereinafter referred to as ES) No mode (schema) Of , When we execute the following command ：

curl -XPUT http://localhost:9200/test/item/1 -d '{"name":"bob", "description": "A Pretty cool guy."}'

ES Can be very smart to identify "name" and "description" The type of field is string, ES The following... Will be created by default mapping：

mappings: {
    item: {
        properties: {
            description: {
                type: string
            }
            name: {
                type: string
            }
        }
    }
}

What is? mapping

ES Of mapping It's very similar to data types in static languages ： Declare a variable as int Variable of type , In the future, this variable can only store int Data of type . alike , One number Type of mapping Fields can only store number Data of type .

Compared to the data type of the language ,mapping There are other implications ,mapping Not only tell ES One field What kind of value is in , It also tells ES How to index data and whether data can be searched .

notes ： When your query does not return the corresponding data , Yours mapping It's very likely that there's a problem . When you're not sure , Check your... Directly mapping.

analyse mapping

One mapping By one or more analyzer form , One analyzer By one or more filter Composed of . When ES When Indexing Documents , It passes the contents of the field to the corresponding analyzer,analyzer And then pass it on to their filters.

filter The function of is easy to understand ： One filter Is a method of converting data , Enter a string , This method returns another string , For example, a method to convert a string to lowercase is filter A good example .

One analyzer Arranged in a sequence filter form , The process of performing the analysis is one by one filter One filter In turn, calls , ES Store and index the final results .

In conclusion , mapping The function of is to execute a series of instructions to convert the input data into searchable index items .

Default analyzer

Back to our example , ES guess description The fields are string type , By default, you create a string Type of mapping, It uses the default global analyzer, default analyzer yes standard analyzer, This standard analyzer There are three filter：token filter, lowercase filter and stop token filter.

We can type... When we make a query _analyze Keyword view analysis process . Use the following instructions to view description Field conversion process ：

curl -X GET "http://localhost:9200/test/_analyze?analyzer=standard&pretty=true" -d "A Pretty cool guy."
{
  "tokens" : [ {
    "token" : "pretty",
    "start_offset" : 2,
    "end_offset" : 8,
    "type" : "<ALPHANUM>",
    "position" : 2
  }, {
    "token" : "cool",
    "start_offset" : 9,
    "end_offset" : 13,
    "type" : "<ALPHANUM>",
    "position" : 3
  }, {
    "token" : "guy",
    "start_offset" : 14,
    "end_offset" : 17,
    "type" : "<ALPHANUM>",
    "position" : 4
  } ]

You can see , our description The value of the field is converted to [pretty], [cool], [guy], Capitalized during conversion A, Punctuation marks are filter Filtered out , Pretty It is also converted to all lowercase pretty, Here's the important thing , Even if ES When storing data, the complete data is still stored , But these three words are the only keywords that can search this data , All the others were abandoned .

Look at the words a To search for results ：

$ curl -X GET "http://localhost:9200/test/_search?pretty=true" -d '{
    "query" : {
        "text" : { "description": "a" }
    }
}'
 
{
  "took" : 29,
  "timed_out" : false,
  "_shards" : {
    "total" : 5,
    "successful" : 5,
    "failed" : 0
  },
  "hits" : {
    "total" : 0,
    "max_score" : null,
    "hits" : [ ]
  }
}

text Type of search uses the same analysis as the previously inserted data in the query process / Filtration system , So we type "a",mapping There will be no return , Because of the word “a” It won't be ES Storage and indexing . In turn, , If we use words "cool" To search ：

curl -X GET "http://localhost:9200/test/_search?pretty=true" -d '{
    "query" : {
        "text" : { "description": "cool" }
    }
}'
 
{
  "took" : 29,
  "timed_out" : false,
  "_shards" : {
    "total" : 5,
    "successful" : 5,
    "failed" : 0
  },
  "hits" : {
    "total" : 1,
    "max_score" : 0.15342641,
    "hits" : [ {
      "_index" : "test",
      "_type" : "item",
      "_id" : "1",
      "_score" : 0.15342641, "_source" : {"name":"zach", "description": "A pretty cool guy"}
    } ]
  }
}

Now you can get the right result , This is an accepted simple example , But it describes ES How it works , Don't put the mapping Think of it as a data type , Think of it as a set of instructions to search for data . If you don't want characters "a" Be deleted , You need to modify your analyzer.