ES|QL: validation doesn't work after FORK

[luigidellaquila]

(on CSV spec dataset)
This query returns a correct validation exception

from airport* |  stats  foo = count_distinct(city_boundary) by type

"type": "verification_exception",
"reason": "Found 1 problem\nline 1:31: argument of [count_distinct(city_boundary)] must be [any exact type except unsigned_long, _source, or counter types], found value [city_boundary] type [geo_shape]"

Adding a FORK, the validation process apparently fails to spot the problem; the result is an esql_illegal_argument_exception (error 500)

from airport* 
| FORK ( WHERE true ) ( WHERE true )  
| WHERE _fork == \"fork0\"
| DROP _fork 
| stats  foo = count_distinct(city_boundary) by type

{
    "error": {
        "root_cause": [
            {
                "type": "esql_illegal_argument_exception",
                "reason": "illegal data type [geo_shape]"
            }
        ],
        "type": "esql_illegal_argument_exception",
        "reason": "illegal data type [geo_shape]"
    },
    "status": 500
}

Relevant stack trace:

ESQL request failed with status [INTERNAL_SERVER_ERROR]: org.elasticsearch.xpack.esql.EsqlIllegalArgumentException: illegal data type [geo_shape]
        at org.elasticsearch.xpack.esql.EsqlIllegalArgumentException.illegalDataType(EsqlIllegalArgumentException.java:43)
        at org.elasticsearch.xpack.esql.EsqlIllegalArgumentException.illegalDataType(EsqlIllegalArgumentException.java:39)
        at org.elasticsearch.xpack.esql.expression.function.aggregate.CountDistinct.supplier(CountDistinct.java:218)
        at org.elasticsearch.xpack.esql.planner.AggregateMapper.entryForAgg(AggregateMapper.java:79)
        at org.elasticsearch.xpack.esql.planner.AggregateMapper.computeEntryForAgg(AggregateMapper.java:67)
        at org.elasticsearch.xpack.esql.planner.AggregateMapper.lambda$map$2(AggregateMapper.java:62)
        at java.base/java.util.HashMap.computeIfAbsent(HashMap.java:1228)
        at org.elasticsearch.xpack.esql.planner.AggregateMapper.map(AggregateMapper.java:62)
        at org.elasticsearch.xpack.esql.planner.AggregateMapper.lambda$doMapping$0(AggregateMapper.java:53)
        at java.base/java.util.stream.ReferencePipeline$7$1.accept(ReferencePipeline.java:273)

[elasticsearchmachine]

Pinging @elastic/es-analytical-engine (Team:Analytics)

[luigidellaquila]

The problem seems to be more general

from airport* 
| FORK ( WHERE true ) ( WHERE true )  
| WHERE _fork == \"fork0\" 
| DROP _fork 
| EVAL a = substring(1, 2, 3)

{
    "error": {
        "root_cause": [
            {
                "type": "class_cast_exception",
                "reason": "class org.elasticsearch.compute.data.IntVectorBlock cannot be cast to class org.elasticsearch.compute.data.BytesRefBlock (org.elasticsearch.compute.data.IntVectorBlock and org.elasticsearch.compute.data.BytesRefBlock are in unnamed module of loader java.net.FactoryURLClassLoader @583fb274)"
            }
        ],
        "type": "class_cast_exception",
        "reason": "class org.elasticsearch.compute.data.IntVectorBlock cannot be cast to class org.elasticsearch.compute.data.BytesRefBlock (org.elasticsearch.compute.data.IntVectorBlock and org.elasticsearch.compute.data.BytesRefBlock are in unnamed module of loader java.net.FactoryURLClassLoader @583fb274)"
    },
    "status": 500
}

[luigidellaquila]

Even more strange, the problem only happens with index patterns and when there are multiple indices involved:

from employees 
| FORK ( WHERE true ) ( WHERE true )  
| WHERE _fork == \"fork0\" 
| DROP _fork 
| EVAL a = substring(1, 2, 3)

{
    "error": {
        "root_cause": [
            {
                "type": "verification_exception",
                "reason": "Found 1 problem\nline 1:103: first argument of [substring(1, 2, 3)] must be [string], found value [1] type [integer]"
            }
        ],
        "type": "verification_exception",
        "reason": "Found 1 problem\nline 1:103: first argument of [substring(1, 2, 3)] must be [string], found value [1] type [integer]"
    },
    "status": 400
}

from employees* 
| FORK ( WHERE true ) ( WHERE true )  
| WHERE _fork == \"fork0\" 
| DROP _fork 
| EVAL a = substring(1, 2, 3)

{
    "error": {
        "root_cause": [
            {
                "type": "class_cast_exception",
                "reason": "class org.elasticsearch.compute.data.IntVectorBlock cannot be cast to class org.elasticsearch.compute.data.BytesRefBlock (org.elasticsearch.compute.data.IntVectorBlock and org.elasticsearch.compute.data.BytesRefBlock are in unnamed module of loader java.net.FactoryURLClassLoader @583fb274)"
            }
        ],
        "type": "class_cast_exception",
        "reason": "class org.elasticsearch.compute.data.IntVectorBlock cannot be cast to class org.elasticsearch.compute.data.BytesRefBlock (org.elasticsearch.compute.data.IntVectorBlock and org.elasticsearch.compute.data.BytesRefBlock are in unnamed module of loader java.net.FactoryURLClassLoader @583fb274)"
    },
    "status": 500
}