Elasticsearch aggregation on different search in same query - elasticsearch

Question

I want to make a query to aggregate base only on match no matter what other parameters(terms , term , etc...) are used.
To be more specific I have an online shop where I use multiple filters (color ,size etc..) If I check a field for example color : red the other colors are no longer aggregated.
A solution that I am using is to make 2 separated queries (one for search where filters are applied and other for aggregation. Any idea how can I combine the 2 separated queries ?

Answer 1

You can take advantage of post_filter which will not apply to your aggregations but will only filter the to-be-returned hits. For example:
Create a shop
PUT online_shop
{
"mappings": {
"properties": {
"color": {
"type": "keyword"
},
"size": {
"type": "integer"
},
"name": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
}
}
}
}
Populate it w/ a few products
POST online_shop/_doc
{"color":"red","size":35,"name":"Louboutin High heels abc"}
POST online_shop/_doc
{"color":"black","size":34,"name":"Louboutin Boots abc"}
POST online_shop/_doc
{"color":"yellow","size":36,"name":"XYZ abc"}
Apply a shared query to the hits as well as aggregations and use post_filter to ... post-filter the hits:
GET online_shop/_search
{
"query": {
"bool": {
"must": [
{
"match": {
"name": "abc"
}
}
]
}
},
"aggs": {
"by_color": {
"terms": {
"field": "color"
}
},
"by_size": {
"terms": {
"field": "size"
}
}
},
"post_filter": {
"bool": {
"must": [
{
"term": {
"color": {
"value": "red"
}
}
}
]
}
}
}
Expected result
{
...
"hits" : {
"total" : {
"value" : 1,
"relation" : "eq"
},
"max_score" : 0.11750763,
"hits" : [
{
"_index" : "online_shop",
"_type" : "_doc",
"_id" : "cehma3IBG_KW3EFn1QYa",
"_score" : 0.11750763,
"_source" : {
"color" : "red",
"size" : 35,
"name" : "Louboutin High heels abc"
}
}
]
},
"aggregations" : {
"by_color" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : "black",
"doc_count" : 1
},
{
"key" : "red",
"doc_count" : 1
},
{
"key" : "yellow",
"doc_count" : 1
}
]
},
"by_size" : {
"doc_count_error_upper_bound" : 0,
"sum_other_doc_count" : 0,
"buckets" : [
{
"key" : 34,
"doc_count" : 1
},
{
"key" : 35,
"doc_count" : 1
},
{
"key" : 36,
"doc_count" : 1
}
]
}
}
}