Define an aggregate rule, and then use its child rules to retrieve the values to aggregate over.
For example, you have some JSON data:
[
{
"name": "Daisy",
"eggs": [ 1, 1, 0, 1, 2, 0, 1 ]
"legs": 2
},
{
"name": "Phoebe",
"eggs": [ 0, 0, 0, 0, 0, 1, 0 ]
"legs": 2
},
{
"name": "Willow",
"eggs": [ 0, 0, 0, 0, 0, 0, 0 ],
"legs": 4
}
]
Some simple rules to generate some aggregate data:
{
"root": "$",
"rules":
[
{
"path": "$",
"target": "creatures", // target not actually used
"interpretation": "IterateListItems",
"children":
[
{
"path": "$.name",
"target": "name",
"interpretation": "AsString"
},
{
"path": "$.legs",
"target": "legs",
"interpretation": "AsNumber"
},
{
"path": "$", // path not actually used
"target": "total-eggs",
"interpretation": "AsAggregateSum",
"children":
[
{
"path": "$",
"target": "eggs", // target not actually used
"interpretation": "IterateListItems",
"children":
[
{
"path": "$",
"target": "eggs-today",
"interpretation": "AsNumber"
}
]
}
]
}
]
}
]
}
In this example, the aggregation is over the number of eggs produced by each creature.
You can see that the total-eggs aggregate value operates over values returned by its child rules.
Those rules do not create columns in the table in their own right - instead they return data which is used by the aggregator to create an aggregate value.
All numerical values returned by the child rules are used for the aggregation.
NB. For this example, the root object is a list. See: Processing a JSON list