Cannot run exists query on field with thousands of nested fields

[merito]

Describe the feature: I've got some documents with field A and inside A are thousands of nested fields. I would like to run exists query to get documents with field A presents

{
  "query": {
    "bool": {
      "must": {
        "exists": {
          "field": "A"
        }
      }
    }
  }
}

but then I've got an error

"type": "too_many_clauses",
"reason": "maxClauseCount is set to 1024"

In my case there is one nested field exists in every A, so I can query for A.B, but what can I do if there is no common fields in A and I want to see all docs with A?

[clintongormley]

Do you mean nested fields or inner object fields? Either way, having thousands of fields is costly, especially if those fields are sparse. We are currently working on adding various soft limits to things like the number of fields you can have in an index, plus the number of nested fields you can have in an index.

Usually your data model can be reworked to use fewer fields by specifying whatever you are using for a field name today as a value instead, eg:

{ "foo": "bar"}

could be better represented as:

{ "key": "foo", "value": bar" }

An alternative, with your current model, is to use a single field within A to identify documents that have an A.* field, eg:

{ "A": {
    "some_val": "foo",
    "exists": true
}}

Then you could just query on A.exists. This will be much more efficient than querying for thousands of potential terms.

[merito]

Yes, I mean inner objects. But still even if there is not thousands of fields I think it will be useful to have a feature to check if only the outer object exists without checking all inner objects, like in my case.

[clintongormley]

@merito see #17186