Percentage calculation in Druid - druid

Question

I have a data in the following format. I am having a hard time doing profit percentage calculation of a particular item or brand w.r.t to all products on a particular day.
Date Item Brand Profit
15-08-2019 A Nike 5
15-08-2019 B Nike 10
15-08-2019 C Nike 12
15-08-2019 D Nike 6
15-08-2019 E Nike 9
15-08-2019 F Adidas 4
15-08-2019 G Adidas 3
15-08-2019 H Adidas 7
16-08-2019 A Nike 8
16-08-2019 B Nike 4
16-08-2019 C Nike 6
16-08-2019 D Nike 7
16-08-2019 E Nike 9
16-08-2019 F Adidas 5
16-08-2019 G Adidas 4
16-08-2019 H Adidas 9
percentage profit of product A on 15th August = profit of A/sum of profits on 15th Aug (5/56).
percentage profit of Nike on 16th Aug = 34/52.
I need to do this calculation in a single query.
If we run it in two parts- first query will fetch data with the date and dimension filters and the second query will fetch me the data with only the date filter. Then i will divide these two.
Cannot figure out a way to combine these in just one query and do the calculation.

Answer 1

Profit percentage calculation of a particular item or brand in your case can be done using filtered aggregation.
For brand 'Nike' following json will work:
{
"queryType": "timeseries",
"dataSource": <your_datasource>,
"granularity": "day",
"aggregations": [
{
"type" : "filtered",
"filter" : {
"type" : "selector",
"dimension" : "Brand",
"value" : "Nike"
},
"aggregator" : {
"type" : "longSum",
"name" : "brand_sum",
"fieldName" : "Profit"
}
},
{ "type": "longSum", "name": "total_sum", "fieldName": "Profit" }
],
"postAggregations": [
{ "type": "arithmetic",
"name": "average_profit",
"fn": "/",
"fields": [
{ "type": "fieldAccess", "name": "brand_sum", "fieldName": "brand_sum" },
{ "type": "fieldAccess", "name": "total_sum", "fieldName": "total_sum" }
]
}
],
"intervals": [ "15-08-2019/16-08-2019" ]
}
It will give average of profit for particular brand (here 'Nike') for all the days (here 15th and 16th.)