Checking if a field exists for any and/or all nested objects - elasticsearch

Question

I took a look at ElasticSearch: search inside the array of objects and while it helps, I'm actually trying to determine if at least one has a field and if all nested objects have the field.
Pretending we have an index of all refrigerators with a superfluous document like:
{
"_id": "whatever",
"location": "North Building 1",
"floor": 2,
"tag": "refrigerator-1",
"contents" : [
{
"item": "milk-carton",
"expires": 1-1-2023
},
{
"item": "pyrex-container",
}
]
}
How do I create an Elastic search query to;
Find any refrigerator that has at least 1 item that CAN expire ( "exists" : { "field" : "expires" } }
Find refrigerators that have no items that expire
Find refrigerators that where all items have an expire field

Answer 1

If you want to do this in a single query , use named_queries
Query
{
"query": {
"bool": {
"should": [
{
"nested": {
"_name": "At least one expires",
"path": "contents",
"query": {
"exists": {
"field": "contents.expires"
}
}
}
},
{
"bool": {
"_name": "None expires",
"must_not": [
{
"nested": {
"path": "contents",
"query": {
"exists": {
"field": "contents.expires"
}
}
}
}
]
}
},
{
"bool": {
"_name": "All expires",
"must": [
{
"nested": {
"path": "contents",
"query": {
"exists": {
"field": "contents.expires"
}
}
}
}
],
"must_not": [
{
"nested": {
"path": "contents",
"query": {
"bool": {
"must_not": [
{
"exists": {
"field": "contents.expires"
}
}
]
}
}
}
}
]
}
}
]
}
}
}
Result
"hits" : [
{
"_index" : "index70",
"_type" : "_doc",
"_id" : "Qt2PVoQB_m3FhzcGBasD",
"_score" : 2.0,
"_source" : {
"location" : "North Building 1",
"floor" : 3,
"tag" : "refrigerator-3",
"contents" : [
{
"item" : "milk-carton",
"expires" : "2023-01-01"
},
{
"item" : "pyrex-container",
"expires" : "2023-01-01"
}
]
},
"matched_queries" : [
"At least one expires",
"All expires"
]
},
{
"_index" : "index70",
"_type" : "_doc",
"_id" : "QN2BVoQB_m3FhzcG9qsG",
"_score" : 1.0,
"_source" : {
"location" : "North Building 1",
"floor" : 2,
"tag" : "refrigerator-1",
"contents" : [
{
"item" : "milk-carton",
"expires" : "2023-01-01"
},
{
"item" : "pyrex-container"
}
]
},
"matched_queries" : [
"At least one expires"
]
},
{
"_index" : "index70",
"_type" : "_doc",
"_id" : "Qd2HVoQB_m3FhzcGUauO",
"_score" : 0.0,
"_source" : {
"location" : "North Building 1",
"floor" : 3,
"tag" : "refrigerator-2",
"contents" : [
{
"item" : "milk-carton"
},
{
"item" : "pyrex-container"
}
]
},
"matched_queries" : [
"None expires"
]
}
]
Query is self explanatory. If you want use separate queries for three conditions, break above query. Each should clause will become a separate query