Elasticsearch. Query for pair of attributes in a nested object

Question

Part of my mapping is:

"individual_attributes" : {
    "type" : "nested",
        "properties" : {
            "template_id" : {"type" : "integer"},
            "attributes_set" : {
                "type" : "nested",
                "properties" : {
                    "attribute_id" : {"type" : "integer"},
                    "attribute_value" : {"type" : "string", "index" : "not_analyzed"}
                }
            }
        }
    }

I need to filter those document which has attribute_id='x' with attribute_value for given id equals 'y'. So that I need matching for pair of fields. Is it possible? Or do I need to change my mapping to something like this:

"individual_attributes" : {
    "type" : "nested",
        "properties" : {
            "template_id" : {"type" : "integer"},
            "attributes_set" : {
                "type" : "nested",
                "properties" : {
                    "attribute_id" : {"type" : "integer",
                        "properties" : {
                            "attribute_value" : {"type" : "string", "index" : "not_analyzed"}
                        }
                    },

                }
            }
        }
    }

Sample data:

                  "attributes_set": [
                 {
                    "attribute_id": 17,
                    "attribute_value": "dolorum"
                 },
                 {
                    "attribute_id": 15,
                    "attribute_value": "at"
                 },
                 {
                    "attribute_id": 18,
                    "attribute_value": "maxime"
                 },
                 {
                    "attribute_id": 14,
                    "attribute_value": "et"
                 },
                 {
                    "attribute_id": 11,
                    "attribute_value": "nemo"
                 },
                 {
                    "attribute_id": 12,
                    "attribute_value": "rem"
                 },
                 {
                    "attribute_id": 10,
                    "attribute_value": "eius"
                 },
                 {
                    "attribute_id": 19,
                    "attribute_value": "deleniti"
                 },
                 {
                    "attribute_id": 13,
                    "attribute_value": "modi"
                 },
                 {
                    "attribute_id": 16,
                    "attribute_value": "neque"
                 }
              ]

I need: SELECT * WHERE attribute(16, for instance)=value(neque). Another words I need matching of pair of fields within one data set:

{
    "attribute_id": x,
    "attribute_value": "y"
}

Sloan Ahrens · Accepted Answer · 2015-10-20 18:18:19Z

2

Here is a simplified example. Your first mapping should be fine for what you want to do. I took out one level of nesting to simplify the explanation, but the same principle will work with arbitrary levels of nesting (if you don't see how to generalize my example I can edit the answer with another example).

I set up a simple mapping like this:

PUT /test_index
{
   "mappings": {
      "doc": {
         "properties": {
            "attributes_set": {
               "type": "nested",
               "properties": {
                  "attribute_id": {
                     "type": "integer"
                  },
                  "attribute_value": {
                     "type": "string",
                     "index": "not_analyzed"
                  }
               }
            }
         }
      }
   }
}

Then added two documents with two nested documents each:

POST /test_index/doc/_bulk
{"index":{"_id":1}}
{"attributes_set": [{"attribute_id": 18,"attribute_value": "dolorum"},{"attribute_id": 15,"attribute_value": "at"}]}
{"index":{"_id":2}}
{"attributes_set": [{"attribute_id": 18,"attribute_value": "maxime"},{"attribute_id": 14,"attribute_value": "et"}]}

Now I can query for a document with a specific nested doc as follows:

POST /test_index/_search
{
   "filter": {
      "nested": {
         "path": "attributes_set",
         "filter": {
            "bool": {
               "must": [
                  {
                     "term": {
                        "attributes_set.attribute_id": {
                           "value": 18
                        }
                     }
                  },
                  {
                     "term": {
                        "attributes_set.attribute_value": {
                           "value": "maxime"
                        }
                     }
                  }
               ]
            }
         }
      }
   }
}

which returns:

{
   "took": 24,
   "timed_out": false,
   "_shards": {
      "total": 5,
      "successful": 5,
      "failed": 0
   },
   "hits": {
      "total": 1,
      "max_score": 1,
      "hits": [
         {
            "_index": "test_index",
            "_type": "doc",
            "_id": "2",
            "_score": 1,
            "_source": {
               "attributes_set": [
                  {
                     "attribute_id": 18,
                     "attribute_value": "maxime"
                  },
                  {
                     "attribute_id": 14,
                     "attribute_value": "et"
                  }
               ]
            }
         }
      ]
   }
}

Here's all the code I used to test it:

http://sense.qbox.io/gist/5e75461a4f0cf96e012cbf0f8262b22f3f8e5ec0

Does that help?

answered Oct 20, 2015 at 18:18

Sloan Ahrens

8,7382 gold badges32 silver badges33 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

John Smith Over a year ago

Thank you VERY much for your time and patience while you have been writing the answer. Very straight and detailed. That's exactly what I need. I've changed your code for using more nesting levels. Becides, I didn't know about sense.qbox.io. Useful instrument. Thank you for pointing it out too. Now it is time to add aggregations to my filter. )

John Smith Over a year ago

And how to add more attribute sets? Tried like that, but it didn't work: "bool": [ {"must": [{"term": {...}}, ...]}, {"must": [{"term": {...}}, ...]} ]

Sloan Ahrens Over a year ago

You have to take the whole nested clause, and put it inside a bool-must, together with another nested clause that would match the second nested document you are looking for. Remember that the nested filter says, "find a nested document that meets these criteria". If you're looking for more than one nested document, you need more than one nested filter.

John Smith Over a year ago

Yes, now it works too. Elasticsearch may have rather big and complex queries... And I start to understand how it works. Thank you.

eemp · Accepted Answer · 2015-10-20 17:59:26Z

1

It looks like you have the perfect mapping in place for this and so all you need to do is build the right query involving a nested query.

Please refer to Elasticsearch docs on how you can query nested objects. See their Querying a Nested Object doc which I think provides all the details you need and a relevant comments example that resembles your case.

EDIT:

Sorry, I just noticed your question regarding the mapping - your first mapping is correct.

answered Oct 20, 2015 at 17:59

eemp

1,1769 silver badges17 bronze badges

Collectives™ on Stack Overflow

Elasticsearch. Query for pair of attributes in a nested object

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related