Elasticsearch Java - Search with should clause

Question

I have a model for Book saved in elastic search. It has couple of attribute minimumAge and maximumAge, which determines what age the book is recommended for

I have couple of usecases for searching booking. First case is to search books by minimumAge. So suppose user searches for minimumAge 5, I have to get all the books which requires minimumAge to be 5 or those books which does not have minimumAge .

I did this using should clause and setting one minimum should clause to satisfy the query which works fine as shown below

private void addSearchByMinimumAge(ProductsFilter filter, BoolQueryBuilder rootQuery) {
    BoolQueryBuilder minAgeQuery = boolQuery();
    minAgeQuery.minimumShouldMatch(1);
    minAgeQuery.should(rangeQuery("minAge").lte(filter.getMinAge()));

    BoolQueryBuilder minAgeNotExistsQuery = boolQuery();
    minAgeNotExistsQuery.mustNot(existsQuery("minimumAge"));
    minAgeQuery.should(minAgeNotExistsQuery);
    rootQuery.must(minAgeQuery);
  }

As mentioned above use case works fine.

Second use case that I am trying to achieve is to get books by age. For example if a person searches for age 5, I need all the books where

• Age 5 is between or is equal to minAge and maxAge.
• If both minAge and maxAge is not available for a book, I should still return the book.
• If minAge is empty for a book but maxAge is less than or equal to 5 then that book should be returned.
• And finally if maxAge is empty then all the books with minAge greater than or equal to 5 should also be returned.

{ "name" : "Book1" , "minAge": 3 , "maxAge": 8 }
{ "name" : "Book2" , "minAge": 6 , "maxAge": 8 }
{ "name" : "Book3" , "minAge": 3 , "maxAge": 5 }
{ "name" : "Book4" , "minAge": 5 , "maxAge": 8 }
{ "name" : "Book5" }
{ "name" : "Book6" , "minAge": 3 }
{ "name" : "Book7" , "minAge": 6 }
{ "name" : "Book8" , "maxAge": 6 }
{ "name" : "Book9" , "maxAge": 4 }

In the above dataset if I pass Books 1,3,4,5,6,8 are valid book results

So I am stuck as to what would be a compact query to address this particular scenario. My first thought process is similar to first query but four should clause and matching any 1. Any pointers will be appreciated

Answer 1

You have to write each case in should clause to satisfy all the scenarios. Below elastic query will cover all your requirements -

{
  "query": {
    "bool": {
      "should": [
        {
          "bool" : {
             "must" : [
               {
                 "range": {
                   "minAge": {
                     "lte": 5
                    }
                  }
               },
               {
                 "range": {
                    "maxAge": {
                      "gte": 5
                     }
                  }
               }
             ] 
          } 
        },
        {
          "bool": {
            "must_not": [
              {
                "exists": {
                  "field": "minAge"
                }
              },
              {
                "exists": {
                  "field": "maxAge"
                }
              }
            ]
          }
        }, 
        {
          "bool" : {
              "must" : [
                {
                 "range": {
                   "minAge": {
                     "lte": 5
                    }
                  }
                },
                {
                  "bool" : {
                     "must_not" : {
                         "exists" : {
                            "field" : "maxAge" 
                          } 
                      } 
                   } 
                } 
              ] 
           } 
        }, 
        {
          "bool" : {
              "must" : [
                {
                 "range": {
                   "maxAge": {
                     "gte": 5
                    }
                  }
                },
                {
                  "bool" : {
                     "must_not" : {
                         "exists" : {
                            "field" : "minAge" 
                          } 
                      } 
                   } 
                } 
              ] 
           } 
        } 
      ]
    }
  }
}