CODEWITHSUNDEEP
elasticsearchelasticsearch-aggregation
Bob Tymczasowy
Bob Tymczasowy

Reputation: 63

Elasticsearch query with grouping

I have database with products. Each Product is composed of fields: uuid, group_id, title, since, till.

since and till define interval of availability.

Intervals [since, till] are disjoint pairs for each group_id. So there are no 2 products within one group for which intervals intersect.

I need to fetch a list of products that meets the following conditions:

  1. on the list should be at most 1 product from each group
  2. each product matches the given title
  3. each product is current (since <= NOW <= till) OR if current product does not exist within its group, it should be the nearest product from the future (min(since) such that since >= NOW)

ES mapping:

{
    "products": {
        "mappings": {
            "dynamic": "false",
            "properties": {
                "group_id": {
                    "type": "long",
                    "fields": {
                        "keyword": {
                            "type": "keyword",
                            "ignore_above": 256
                        }
                    }
                },
                "title": {
                    "type": "text",
                    "fields": {
                        "keyword": {
                            "type": "keyword",
                            "ignore_above": 256
                        }
                    }
                },
                "since": {
                    "type": "date",
                    "fields": {
                        "keyword": {
                            "type": "keyword",
                            "ignore_above": 256
                        }
                    }
                },
                "till": {
                    "type": "date",
                    "fields": {
                        "keyword": {
                            "type": "keyword",
                            "ignore_above": 256
                        }
                    }
                }
            }
        }
    }
}

Is it possible to create such query in Elasticsearch?

Upvotes: 2

Views: 3777

Answers (1)

Kamal Kunjapur
Kamal Kunjapur

Reputation: 8840

Looking at your mapping, I've created sample documents, the query and its response as below:

Sample Documents:

POST product_index/_doc/1
{
  "group_id": 1,
  "title": "nike",
  "since": "2020-01-01",
  "till": "2020-03-31"
}

POST product_index/_doc/2
{
  "group_id": 2,
  "title": "nike",
  "since": "2020-01-01",
  "till": "2020-03-31"
}

POST product_index/_doc/3
{
  "group_id": 3,
  "title": "nike",
  "since": "2020-03-15",
  "till": "2020-03-31"
}

POST product_index/_doc/4
{
  "group_id": 3,
  "title": "nike",
  "since": "2020-03-19",
  "till": "2020-03-31"
}

As mentioned above, there are like 4 documents in total, group 1 and 2 have one document each while group 3 has two documents with both since >= now

Query Request:

The summary of the query is below:

Bool
 - Must 
   - Match title as nike 
   - Should 
     - clause 1 - since <= now <= till
     - clause 2 - now <= since
Agg
 - Terms on GroupId
   - Top Hits (retrieve only 1st document as your clause is at most for each group, and sort them by asc order of since)

Below is the actual query:

POST product_index/_search
{
  "size": 0,
  "query": {
    "bool": {
      "must": [
        {
          "match": {
            "title": "nike"
          }
        },
        {
          "bool": {
            "should": [
              {                               <--- since <=now <= till
                "bool": {
                  "must": [
                    {
                      "range": {
                        "till": {
                          "gte": "now"
                        }
                      }
                    },
                    {
                      "range": {
                        "since": {
                          "lte": "now"
                        }
                      }
                    }
                  ]
                }
              },
              {                               <---- since >= now
                "bool": {
                  "must": [
                    {
                      "range": {
                        "since": {
                          "gte": "now"
                        }
                      }
                    }
                  ]
                }
              }
            ]
          }
        }
      ]
    }
  },
  "aggs": {
    "my_groups": {
      "terms": {
        "field": "group_id.keyword",
        "size": 10
      },
      "aggs": {
        "my_docs": {
          "top_hits": {
            "size": 1,                           <--- Note this to return at most one document
            "sort": [
              { "since": { "order": "asc"}       <--- Sort to return the lowest value of since
              }
             ]  
          }
        }
      }
    }
  }
}

Notice that I've made use of Terms Aggregation and Top Hits as its sub-aggregation.

Response:

{
  "took" : 7,
  "timed_out" : false,
  "_shards" : {
    "total" : 1,
    "successful" : 1,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 4,
      "relation" : "eq"
    },
    "max_score" : null,
    "hits" : [ ]
  },
  "aggregations" : {
    "my_groups" : {
      "doc_count_error_upper_bound" : 0,
      "sum_other_doc_count" : 0,
      "buckets" : [
        {
          "key" : "3",
          "doc_count" : 2,
          "my_docs" : {
            "hits" : {
              "total" : {
                "value" : 2,
                "relation" : "eq"
              },
              "max_score" : null,
              "hits" : [
                {
                  "_index" : "product_index",
                  "_type" : "_doc",
                  "_id" : "3",
                  "_score" : null,
                  "_source" : {
                    "group_id" : 3,
                    "title" : "nike",
                    "since" : "2020-03-15",
                    "till" : "2020-03-31"
                  },
                  "sort" : [
                    1584230400000
                  ]
                }
              ]
            }
          }
        },
        {
          "key" : "1",
          "doc_count" : 1,
          "my_docs" : {
            "hits" : {
              "total" : {
                "value" : 1,
                "relation" : "eq"
              },
              "max_score" : null,
              "hits" : [
                {
                  "_index" : "product_index",
                  "_type" : "_doc",
                  "_id" : "1",
                  "_score" : null,
                  "_source" : {
                    "group_id" : 1,
                    "title" : "nike",
                    "since" : "2020-01-01",
                    "till" : "2020-03-31"
                  },
                  "sort" : [
                    1577836800000
                  ]
                }
              ]
            }
          }
        },
        {
          "key" : "2",
          "doc_count" : 1,
          "my_docs" : {
            "hits" : {
              "total" : {
                "value" : 1,
                "relation" : "eq"
              },
              "max_score" : null,
              "hits" : [
                {
                  "_index" : "product_index",
                  "_type" : "_doc",
                  "_id" : "2",
                  "_score" : null,
                  "_source" : {
                    "group_id" : 2,
                    "title" : "nike",
                    "since" : "2020-01-01",
                    "till" : "2020-03-31"
                  },
                  "sort" : [
                    1577836800000
                  ]
                }
              ]
            }
          }
        }
      ]
    }
  }
}

Let me know if this helps!

Upvotes: 5

Related Questions