Search Service aggregate: how to prevent summarizing small buckets (sum_other_doc_count)

Question

Hi all,

do you know how i can prevent to sum up "sum_other_doc_count" in response of a search service/aggregate request? I want to have listed all buckets which have at least a doc_count >=1

i tried to search like this:

"aggs":{
  "anyObject.status": {
  	"sub":{
  		"anyObject.segment":{
  		}
  	}
        }
  }

and get this as response:

i already searche in elastic documentation but couldn't find a solution...

Many thx in advance
Martin

Answer 1

Hi Martin,

I might need more information to help you. How many buckets did you get?

Best regards
Irina

Comments

Hi Irina,

i'll get 10 Buckets for anyObject.Status and for each of theses buckets at least one to a maximum of ten subbuckets...

Thx & Rgds
Martin

---

Hi Martin,

By default, Elasticsearch will return the buckets for the top ten terms ordered by the doc_count.
We use the default behavior because the higher the requested size is, the more accurate the results will be, but also, the more expensive it will be to compute the final results.

To solve your problem we could extend our query and  provide size of term buckets that should be returned out of the overall terms list.  It will be a new feature.

Best regards
Irina

---

Hi Irina,
thx, would be great if you could extend your query...
You're talking about  a new feature, does this mean that there will be no fix or update for 5.20 lts?

Thx & Rgds
Martin

---

Hi Martin,

yes, this new feature will not come for the 5.20 lts.

Please, can you give me a bit more details why you need this feature?

Best Regards
Martin

---

Hi Martin,

thx for the information.
We have a specific Use case which requires the aggragation of all values a defined Object could store.
We developed our own kanban app in Yuuvis, which represents various states of Tasks... to analyse this date we need a daily "snapshot" of alle existing states, stored in mentioned objects... with this information we can render a cumulative-flow-diagram to measure and analyse our processes... but this makes only sense if we get all states and not only top ten :-)
i've seen that bucket sums (analogues to elastic) are also not possible, maybe this could be also feature in upcoming versions? This will be necessary if we are not only counting docs per Bucket in future, but additionaly sum all estimated efforts (maybe hours or days) stored at each task...