Sorting aggregated data in elastic search - elasticsearch

Question

I am doing a search that is doing an aggregation by xyz field and getting the latest version. Now I need to sort the aggregated data based on created field. Let me know how we can do that.
{
"query": {
"query_string": {
"query": ""
}
},
"aggs": {
"uuid": {
"terms": {
"field": "xyz.keyword"
},
"aggs": {
"top_trades_hits": {
"top_hits": {
"sort": [
{
"version": {
"order": "desc"
}
}
],
"size": 1
}
}
}
}
}
}
the Above mentioned query returns
{
"aggregations": {
"uuid": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": "794a5b8f-3e22-4ff9-98bb-b8b54c85948e",
"doc_count": 3,
"agg": {
"hits": {
"total": {
"value": 3,
"relation": "eq"
},
"max_score": null,
"hits": [
{
"_index": "index",
"_type": "doc",
"_id": "7",
"_score": null,
"_source": {
"uuid": "794a5b8f-3e22-4ff9-98bb-b8b54c85948e",
"type": "qsdn",
"discontinued": false,
"minSupportedPlatformVersion": "11.5.3.3",
"version": 2,
"created": 1658428291346
},
"sort": [
2
]
}
]
}
}
},
{
"key": "03504029-a029-417d-bd67-fb1b5fc5055b",
"doc_count": 2,
"agg": {
"hits": {
"total": {
"value": 2,
"relation": "eq"
},
"max_score": null,
"hits": [
{
"_index": "index",
"_type": "doc",
"_id": "9",
"_score": null,
"_source": {
"uuid": "03504029-a029-417d-bd67-fb1b5fc5055b",
"type": "gdsg",
"discontinued": false,
"version": 1.1,
"created": 1554904300799
},
"sort": [
1.1
]
}
]
}
}
}
]
}
}
}
Document for the elastic search is as follows
{
"_index": "index",
"_type": "doc",
"_id": "3",
"_version": 2,
"_seq_no": 1,
"_primary_term": 1,
"found": true,
"_source": {
"doc": {
"uuid": "abcd",
"type": "strifn",
"name": "default",
"version": 3.12,
"s3ObjectVersionId": "",
"created": 165842829134
}
}
}
Expected result
{
"aggregations": {
"uuid": {
"doc_count_error_upper_bound": 0,
"sum_other_doc_count": 0,
"buckets": [
{
"key": "03504029-a029-417d-bd67-fb1b5fc5055b",
"doc_count": 2,
"agg": {
"hits": {
"total": {
"value": 2,
"relation": "eq"
},
"max_score": null,
"hits": [
{
"_index": "index",
"_type": "doc",
"_id": "9",
"_score": null,
"_source": {
"uuid": "03504029-a029-417d-bd67-fb1b5fc5055b",
"type": "gdsg",
"discontinued": false,
"version": 1.1,
"created": 1554904300799
},
"sort": [
1.1
]
}
]
}
}
},
{
"key": "794a5b8f-3e22-4ff9-98bb-b8b54c85948e",
"doc_count": 3,
"agg": {
"hits": {
"total": {
"value": 3,
"relation": "eq"
},
"max_score": null,
"hits": [
{
"_index": "index",
"_type": "doc",
"_id": "7",
"_score": null,
"_source": {
"uuid": "794a5b8f-3e22-4ff9-98bb-b8b54c85948e",
"type": "qsdn",
"discontinued": false,
"minSupportedPlatformVersion": "11.5.3.3",
"version": 2,
"created": 1658428291346
},
"sort": [
2
]
}
]
}
}
}
]
}
}
}
I am using AWS opensearch for the same

Answer 1

Your query is correct only, you just need to increase the size from 1 to see all the documents in your bucket sorted according to version field in your Elasticsearch index.
Can you share more info, if above doesn't help you, like sample documents and index mapping.