Wednesday 21 October 2015

Elasticsearch: mapping

Till now we are posting data to Easticsearch, and not worrying about how each field is mapped to specific data type. Elastic search generates dynamic mapping for us.
For example
GET employees/employee/1
Above query returns following data.

{
   "_index": "employees",
   "_type": "employee",
   "_id": "1",
   "_version": 1,
   "found": true,
   "_source": {
      "firstName": "Tushar",
      "lastName": "Goyal",
      "hobbies": [
         "Reading Novels",
         "Playing football"
      ],
      "age": 23,
      "joiningDate": {
         "year": 2014,
         "month": 1,
         "day": 12
      }
   }
}
You can see mapping of any type using _mapping end point.

GET employees/_mapping/employee
Above statement return mapping for type employee.

{
   "employees": {
      "mappings": {
         "employee": {
            "properties": {
               "age": {
                  "type": "long"
               },
               "firstName": {
                  "type": "string"
               },
               "hobbies": {
                  "type": "string"
               },
               "joiningDate": {
                  "properties": {
                     "day": {
                        "type": "long"
                     },
                     "month": {
                        "type": "long"
                     },
                     "year": {
                        "type": "long"
                     }
                  }
               },
               "lastName": {
                  "type": "string"
               }
            }
         }
      }
   }
}


As you observe all the integers are mapped to long type, and all character sequences are mapped to string type. Long type data represents exact data, string data represent full text data.

Data in Elasticsearch is categorized into two types.
a.   Exact Data
b.   Full text data

a. Exact Data
Exact data is some thing like integers, long, float, date kind of data.

b. Full text data
It refers to textual data. For example news articles, body of E-mail etc., Full text is different from exact search. In exact search we can search for exact values like get me all employees whose salary is > 25000, Get all employees whose age is > 24, get me all students whose DOB is 02-05-1989 etc.,

In case of full text search, we check how well this document matches to given query. For example, when you see any news application, they suggest related articles for given news article. They use full text search to figure out related documents. We will discuss about Full text search in detail later.





Prevoius                                                 Next                                                 Home

No comments:

Post a Comment