Elasticsearch arrays query/filter

Question

I'm looking at Elasticsearch for the first time and spent around a day looking at it. We already use Lucene extensively and want to start using ES instead. I'm looking at alternative data structures to what we currently have.

If I run *match_all* query this is what I get at the moment. I am happy with this structure.

{
   "took": 2,
   "timed_out": false,
   "_shards": {
      "total": 5,
      "successful": 5,
      "failed": 0
   },
   "hits": {
      "total": 22,
      "max_score": 1,
      "hits": [
         {
            "_index": "integration-test-static",
            "_type": "sport",
            "_id": "4d38e07b-f3d3-4af2-9221-60450b18264a",
            "_score": 1,
            "_source": {
               "Descriptions": [
                  {
                     "FeedSource": "dde58b3b-145b-4864-9f7c-43c64c2fe815",
                     "Value": "Football"
                  },
                  {
                     "FeedSource": "e4b9ad44-00d7-4216-adf5-3a37eafc4c93",
                     "Value": "Football"
                  }
               ],
               "Synonyms": [
                  "Football"
               ]
            }
         }
      ]
   }
}

What I can't figure out is how a query is written to pull back this document by searching for the synonym "Football". Looks like it should be easy!

I got this approach after reading this: http://gibrown.wordpress.com/2013/01/24/elasticsearch-five-things-i-was-doing-wrong/ He mentions storing multiple fields in arrays. I realise my example does not have multiple fields, but we will certainly be looking for a solution which can cater for them.

Tried various different queries with filters, bool things, term this and terms that, none return.

Alex Brasetvik · Accepted Answer · 2014-02-20 16:58:05Z

5

What does your search and mappings look like?

If you let Elasticsearch generate the mapping, it'll use the standard analyzer which lowercases the text (and removes stopwords).

So Football will actually be indexed as football. The term-family of queries/filters do not do text analysis, so term:Football will be looking for Football, which is not indexed. The match-family of queries do.

This is a very common problem, and is covered quite extensively in my article on Troubleshooting Elasticsearch searches, for Beginners, which can be worth skimming through. Text analysis is a very important part of working with search, so there's some more articles about it as well.

answered Feb 20, 2014 at 16:58

Alex Brasetvik

11.8k2 gold badges41 silver badges38 bronze badges

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

1 Comment

Matt Canty Over a year ago

Thanks for a prompt reply (and Akshay). I will certainly look through the beginners guide. A few days and it'll all start sinking in...

Akshay · Accepted Answer · 2014-02-20 16:53:41Z

2

A simple match query would work in this scenario.

POST integration-test-static/_search
{
    "query": {
        "match": {
           "Synonyms": "Football"
        }
    }
}

Which returns:

{
   "took": 0,
   "timed_out": false,
   "_shards": {
      "total": 1,
      "successful": 1,
      "failed": 0
   },
   "hits": {
      "total": 1,
      "max_score": 0.30685282,
      "hits": [
         {
            "_index": "integration-test-static",
            "_type": "sport",
            "_id": "4d38e07b-f3d3-4af2-9221-60450b18264a",
            "_score": 0.30685282,
            "_source": {
               "Descriptions": [
                  {
                     "FeedSource": "dde58b3b-145b-4864-9f7c-43c64c2fe815",
                     "Value": "Football"
                  },
                  {
                     "FeedSource": "e4b9ad44-00d7-4216-adf5-3a37eafc4c93",
                     "Value": "Football"
                  }
               ],
               "Synonyms": [
                  "Football"
               ]
            }
         }
      ]
   }
}

answered Feb 20, 2014 at 16:53

Akshay

3,4111 gold badge24 silver badges19 bronze badges

3 Comments

Matt Canty Over a year ago

Thanks I also confirm that this works. Could you possibly show how I might find "Football" only. To avoid returning "American Football" for example?

Akshay Over a year ago

As Alex says, there's a lot you can do with analyzers to help you with your indexing and queries. A simple one to look at is the "keyword" analyzer. That will convert everything to lower case, but treat the entire string as a single word and not split it up. So a search for the term:football will not match term:"american football"

Matt Canty Over a year ago

The keyword analyzer was exactly what I was looking for here.

Collectives™ on Stack Overflow

Elasticsearch arrays query/filter

2 Answers 2

1 Comment

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related