Elasticsearch: Ordering with respect to Nested Sum Aggregation - sorting

Question

The use case is to create aggregations to find top selling products in a region sorted by the counts of products sold. The data is stored in an index in elasticsearch.
I want to sort my 'group_by_name' aggregations output by the sum_quantity value aggregation which is in the last nested aggregation/ two levels in after 'group_by_name' with an intermediate aggregation 'group_by_sku'. The default output is sorted by doc_count. I want the aggregation to be sorted by the 'sum_quantity' aggregation value.
I have an index with the following mapping:
{"settings": {
"index": {
"number_of_shards": 2,
"number_of_replicas": 0
},
"analysis":{
"analyzer":{
"autocomplete":{
"type":"custom",
"tokenizer":"standard",
"filter":[ "standard", "lowercase", "ngram" ]
}
},
"filter":{
"ngram":{
"type":"ngram",
"min_gram":3,
"max_gram":25
}
}
} },"mappings": {
"farmer_products_map":{
"properties": {
"state": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
},
"district": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
},
"taluka": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
},
"village": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
},
"createdOn": {
"type": "date",
"format": "epoch_millis"
},
"category": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
},
"brand": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
},
"productName": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
},
"autocomplete":{
"analyzer": "autocomplete",
"type": "string"
}
}
},
"crop": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
},
"sku": {
"type": "text",
"fields": {
"keyword": {
"type": "keyword"
}
}
},
"quantity": {
"type": "integer"
},
"farmerId": {
"type": "integer"
},
"orderId": {
"type": "integer"
}
}} }}
The following are the sample documents:
[{
"_index": "farmer_products_index_adv",
"_type": "farmer_products_map",
"_id": "AVtCttf0IP9v8cUTtoiz",
"_score": 1,
"_source": {
"orderId": 469173,
"category": "Hardware",
"farmerId": 509583,
"district": "",
"brand": "Honda",
"taluka": "",
"crop": "",
"productName": "Honda BRUSH CUTTER UMK 435-T U2NT",
"state": "",
"sku": "AGS-HW-471",
"village": "",
"quantity": 1
},{
"_index": "farmer_products_index_adv",
"_type": "farmer_products_map",
"_id": "AVtCttf0IP9v8cUTtoi1",
"_score": 1,
"_source": {
"orderId": 469177,
"category": "Crop Nutrition",
"farmerId": 13732,
"district": "Banaskantha",
"brand": "Unassigned Brand",
"taluka": "Kankrej",
"crop": "",
"productName": "Free Power Gel - Plant Nutrient (500 Ml)",
"state": "Gujarat",
"sku": "AGS-CN-006",
"village": "Nanota",
"quantity": 1
}}]
I wish to perform the following query aggregation:
{
"query": {
"bool": {
"must": [{
"match": {
"state": {
"query": "Maharashtra",
"fuzziness": 3,
"prefix_length": 2
}
}
}, {
"match": {
"district": {
"query": "Wardha",
"fuzziness": 3,
"prefix_length": 2
}
}
}, {
"match": {
"taluka": {
"query": "Wardha",
"fuzziness": 3,
"prefix_length": 2
}
}
}]
}
},
"size": 0,
"aggs": {
"group_by_state": {
"terms": {
"field": "state.keyword"
},
"aggs": {
"group_by_district": {
"terms": {
"field": "district.keyword"
},
"aggs": {
"group_by_taluka": {
"terms": {
"field": "taluka.keyword"
},
"aggs": {
"group_by_name": {
"terms": {
"field": "productName.keyword"
},
"aggs": {
"group_by_sku": {
"terms": {
"field": "sku.keyword"
},
"aggs": {
"sum_quantity": {
"sum": {
"field": "quantity"
}
}
}
}
}
}
}
}
}
}
}
}
}}
The current output of the aggregation is:
[{
"key": "Free MH HDPE Tarpaulin Tape Black 3mtr roll",
"doc_count": 13,
"group_by_sku": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [{
"key": "AGS-HW-410",
"doc_count": 13,
"sum_quantity": {
"value": 13
}
}]}}, {
"key": "Tarpaulin Sheet 11*15 (Tadpatri) 250 GSM",
"doc_count": 10,
"group_by_sku": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [{
"key": "AGS-HW-326",
"doc_count": 10,
"sum_quantity": {
"value": 10
}
}]
}}, {
"key": "Free Humic power Advanced powder 95% (250 Gms)",
"doc_count": 6,
"group_by_sku": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [{
"key": "AGS-CN-036",
"doc_count": 6,
"sum_quantity": {
"value": 18
}
}]
}}]
I want the output to be sorted by the sum_quantity value:
[{
"key": "Free Humic power Advanced powder 95% (250 Gms)",
"doc_count": 6,
"group_by_sku": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [{
"key": "AGS-CN-036",
"doc_count": 6,
"sum_quantity": {
"value": 18
}
}]}}, {
"key": "Free MH HDPE Tarpaulin Tape Black 3mtr roll",
"doc_count": 13,
"group_by_sku": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [{
"key": "AGS-HW-410",
"doc_count": 13,
"sum_quantity": {
"value": 13
}
}]
}}, {
"key": "Tarpaulin Sheet 11*15 (Tadpatri) 250 GSM",
"doc_count": 10,
"group_by_sku": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [{
"key": "AGS-HW-326",
"doc_count": 10,
"sum_quantity": {
"value": 10
}
}]
}}]
How can I achieve this? I tried approaches suggested in other stackoverflow questions using "reverse_nested" but was unable to arrive at a solution.