I'm looking at using schema to document the Query and its responses. This is mostly OK, but I've run into a blocker with the Aggregate.

The spec has the following structure:

Aggregate (OPTIONAL)

• Columns (list)
  • ColName: column name, see below
  • Aggregator: aggregator function, one of AVG,
    SUM, COUNT, MIN, MAX, STDDEV
• GroupBy (list) (OPTIONAL)
  --ColName: column name

That suggests an object containing two lists:

{
"Aggregate":

{ "Columns": [], "GroupBy": [] }

}

which could expand to something like this: (A)

{
"Aggregate": {
"Columns": [

{ "ColName": "a", "Aggregator": "SUM" }

,

{ "ColName": "b", "Aggregator": "AVG" }

],
"GroupBy": ["a", "b"]
}
}

of if GroupBy has to have a ColName then it would be this: (B)

{
"Aggregate": {
"Columns": [

{ "ColName": "a", "Aggregator": "SUM" }

,

{ "ColName": "b", "Aggregator": "AVG" }

],
"GroupBy": [

{ "ColName": "a" }

,

{ "ColName": "b" }

]
}
}

We have no examples of GroupBy and the only example of Aggregate (C, below) does not have an array, just a single item:

{
"Aggregate": {
"Columns":

{ "ColName": "WHAT_CLUSTER", "Aggregator": "COUNT" }

}
}