Count unique fields with sum parameter range

[Smasell (Sergey)]

Hi!!!
I have my data points where I have two fields:
user_id
buy_price

Each user_id can sent several data points(with different buy_price value).
Could you please help to count user_id's, grouped by sum their buy_prices(to 100, from 100 to 200, from 200).

Thanks for helping me!

[mayya (Mayya Sharipova)]

If I understood your problem correctly, you wanted to have buckets based on buy_price and for each bucket count of user_id for this price.
Range aggregation should help you here.

GET /_search
{
    "aggs" : {
        "price_ranges" : {
            "range" : {
                "field" : "price",
                "ranges" : [
                    { "to" : 100.0 },
                    { "from" : 100.0, "to" : 200.0 },
                    { "from" : 200.0 }
                ]
            }
        }
    }
}

If this is not what you intended, please provide an example for the response you are looking for.

[Smasell (Sergey)]

Thank's for your response!
Everything right, but it's not what I'm looking for. For example I have several docs:

user_id: Ben buy_price: 50
user_id: Ben buy_price: 100
user_id: Ben buy_price: 100
user_id: Jane buy_price: 50
user_id: Luke buy_price: 250

I need to count sum of their buy_price fields and count how many unique user_id each sum range has.
Response in this case should look like this:

to 100 - 1 (Jane. because sum of all her buys is equal 50)
from 100 to 200 - 0
from 200 - 2(Luke has one doc with price 250 and Ben has 3 docs (100+100+50))

Thanks for helping me.

[mayya (Mayya Sharipova)]

I understand now what you intended to do. Unfortunately, you can not do this in elasticsearch.
This would involve having a pipeline range aggregation, which we have decided not to develop. You can read more here https://github.com/elastic/elasticsearch/issues/17590

[Smasell (Sergey)]

Get it! Mayya, many thanks and good luck!!!)