Elasticsearch on fashion commerce - node.js

Question

I have the following dataset indexed in elasticsearch
[
{
"id": "1",
"name": "Red T-shirt",
"size": "S",
"styleId": "R-1"
},
{
"id": "2",
"name": "Red T-shirt",
"size": "M",
"styleId": "R-1"
},
{
"id": "3",
"name": "Red T-shirt",
"size": "L",
"styleId": "R-1"
},
{
"id": "4",
"name": "White T-shirt",
"size": "S",
"styleId": "W-1"
},
{
"id": "5",
"name": "White T-shirt",
"size": "XL",
"styleId": "W-1"
}
]
In search result I want products grouped by styleId. Is it possible like bellow sample result? In most of the commerce websites like Amazon, Myntra, Flipkart I have seen they are doing the same what I am trying. I have tried many elasticsearch concepts like "collapse" but without success. Using kibana, and elasticsearch version 7.4. Any suggestions will be highly appreciated.
[
{
"id": "1",
"name": "Red T-shirt",
"size": "S",
"styleId": "R-1",
"sizes": [
{
"id": "2",
"name": "Red T-shirt",
"size": "M",
"styleId": "R-1"
},
{
"id": "3",
"name": "Red T-shirt",
"size": "L",
"styleId": "R-1"
}
]
},
{
"id": "4",
"name": "White T-shirt",
"size": "S",
"styleId": "W-1",
"sizes": [
{
"id": "5",
"name": "White T-shirt",
"size": "XL",
"styleId": "W-1"
}
]
},
]

Answer 1

You can use collapse with cardinality aggregation.
Cardinality will give group level count
{
"_source": "false",
"collapse": {
"field": "styleId.keyword",
"inner_hits": {
"name": "group-details",
"size": 5
}
},
"from": 0,
"size": 5,
"aggs": {
"type_count": {
"cardinality": {
"field": "styleId.keyword"
}
}
}
}