# How dfs\_query\_then\_fetch works? **URL:** \ **Category:** Elasticsearch\ **Created:** [January 11, 2018, 3:31am UTC](https://discuss.elastic.co/t/how-dfs-query-then-fetch-works/114995 "2018-01-11T03:31:45Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![ankur\_singla](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/ankur_singla/32/45612_2.png) [@ankur\_singla](https://discuss.elastic.co/u/ankur_singla)\ **Post date:** [January 11, 2018, 3:31am UTC](https://discuss.elastic.co/t/how-dfs-query-then-fetch-works/114995/1 "2018-01-11T03:31:46Z")
Hi guys As per my understanding **dfs\_query\_then\_fetch** is used to get accurate results over multiple shards. **Term aggregation** returns the approximate count per term(still confused with how it calculates documents errors) . If i use dfs\_query\_then\_fetch search with term aggregation will it return the accurate results? I tried but it is not. I am still getting doc\_count\_error\_upper\_bound value greater than zero. **QUERY** : ``` GET demo/_search?pretty=true&search_type=dfs_query_then_fetch { "aggs" : { "products" : { "terms" : { "field" : "demo1.keyword", "size" : 5 } } } } ``` **Result** : ``` "aggregations": { "products": { "doc_count_error_upper_bound": 10959, "sum_other_doc_count": 10454016, "buckets": [ { "key": "SK148", "doc_count": 4442 }, { "key": "SK67", "doc_count": 4432 }, { "key": "SK489", "doc_count": 4420 }, { "key": "SK592", "doc_count": 2245 }, { "key": "SK88", "doc_count": 2245 } ] } } ``` I think i misconception here. Please guide me on this. It will be great help. ---
**Author:** ![Mikhail\_Khludnev](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/mikhail_khludnev/32/59591_2.png) [@Mikhail\_Khludnev](https://discuss.elastic.co/u/Mikhail_Khludnev)\ **Post date:** [January 11, 2018, 4:13am UTC](https://discuss.elastic.co/t/how-dfs-query-then-fetch-works/114995/2 "2018-01-11T04:13:10Z")
I believe _dfs\_query\_then\_fetch_ calculates terms' IDFs globally without per-shard skew. ---
**Author:** ![ankur\_singla](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/ankur_singla/32/45612_2.png) [@ankur\_singla](https://discuss.elastic.co/u/ankur_singla)\ **Post date:** [January 11, 2018, 8:16am UTC](https://discuss.elastic.co/t/how-dfs-query-then-fetch-works/114995/3 "2018-01-11T08:16:14Z")
Yeah. I am assuming this means it will calculate count on whole data rather than on shard level & i should get accurate count without any **doc\_count\_error\_upper\_bound** or zero. Am i right? ---
**Author:** ![Mikhail\_Khludnev](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/mikhail_khludnev/32/59591_2.png) [@Mikhail\_Khludnev](https://discuss.elastic.co/u/Mikhail_Khludnev)\ **Post date:** [January 11, 2018, 1:37pm UTC](https://discuss.elastic.co/t/how-dfs-query-then-fetch-works/114995/4 "2018-01-11T13:37:41Z")
No. I believe dfs\_query\_then\_fetch impacts only scoring (search), but not aggregations. ---
**Author:** ![Madhusudan\_Atmakuri](https://avatars.discourse-cdn.com/v4/letter/m/b4bc9f/32.png) [@Madhusudan\_Atmakuri](https://discuss.elastic.co/u/Madhusudan_Atmakuri)\ **Post date:** [January 11, 2018, 5:19pm UTC](https://discuss.elastic.co/t/how-dfs-query-then-fetch-works/114995/5 "2018-01-11T17:19:35Z")
I am not sure I understand. Based on this: > dfs\_query\_then\_fetch should return result skipping the first phase of query\_then\_fetch. The documentation does not say that it does not impact aggregations. ---
**Author:** ![Mikhail\_Khludnev](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/mikhail_khludnev/32/59591_2.png) [@Mikhail\_Khludnev](https://discuss.elastic.co/u/Mikhail_Khludnev)\ **Post date:** [January 12, 2018, 3:36am UTC](https://discuss.elastic.co/t/how-dfs-query-then-fetch-works/114995/6 "2018-01-12T03:36:50Z")
This issue is about incorrect explanation. Which is the a separate thing. > Madhusudan\_Atmakuri [https://discuss.elastic.co/u/madhusudan\_atmakuri](https://discuss.elastic.co/u/madhusudan_atmakuri) > January 11 > > I am not sure I understand. Based on this: > [GitHub - elastic/elasticsearch: Free and Open, Distributed, RESTful Search Engine](http://github.com/elastic/elasticsearch) > [https://github.com/elastic/elasticsearch/issues/20580](https://github.com/elastic/elasticsearch/issues/20580) > [https://github.com/XGiton](https://github.com/XGiton) Issue: dfs\_query\_then\_fetch does work > sometimes [https://github.com/elastic/elasticsearch/issues/20580](https://github.com/elastic/elasticsearch/issues/20580) > opened by XGiton [https://github.com/XGiton](https://github.com/XGiton) on 2016-09-20 > [https://github.com/elastic/elasticsearch/issues/20580](https://github.com/elastic/elasticsearch/issues/20580) > closed by jimczi [https://github.com/jimczi](https://github.com/jimczi) on 2016-09-20 > [https://github.com/elastic/elasticsearch/issues/20580](https://github.com/elastic/elasticsearch/issues/20580) > > Elasticsearch version: Elasticsearch 5.0.0-alpha5 > Plugins installed: [analyzer-ik] > JVM version: jdk1.8.0\_101 > OS version: Ubuntu 14.04 > Description of the problem including expected versus actual behavior: > dfs\_query\_then\_fetch does not... > > dfs\_query\_then\_fetch should return result skipping the first phase of > query\_then\_fetch. The documentation does not say that it does not impact > aggregations. The documentation doesn't say it does. ---
**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\ **Post date:** [February 9, 2018, 3:37am UTC](https://discuss.elastic.co/t/how-dfs-query-then-fetch-works/114995/7 "2018-02-09T03:37:05Z")
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.