This question is not how to do aggregation by multiple fields, which we can use sub aggregation.
If you know SQL, I can give you a perfect explanation:
SELECT SUM(SomeField1), MAX(SomeField2), MIN(SomeField3), AVG(SomeField4) FROM FooTable
GROUP BY Key1, Key2, Key3
Can we achieve that in Elasticsearch?
Thanks.
Yes, you need to aggregate by Key1 and then add two sub-aggregations for Key2 and Key3 and finally add one metric sub-aggregation per SomeField*, basically like this:
{
"size": 0,
"aggs": {
"key1": {
"terms": {
"field": "Key1"
},
"aggs": {
"key2": {
"terms": {
"field": "Key2"
},
"aggs": {
"key3": {
"terms": {
"field": "Key3"
},
"aggs": {
"somefield1": {
"sum": {
"field": "SomeField1"
}
},
"somefield2": {
"max": {
"field": "SomeField2"
}
},
"somefield3": {
"min": {
"field": "SomeField3"
}
},
"somefield4": {
"avg": {
"field": "SomeField4"
}
}
}
}
}
}
}
}
}
}
