FOREACH func in DQL (loops in Bulk Upsert)

Dev
,
4

Foreach In Query itself

{
	var (func: eq(name, "nodeA")) {
		ED as edge
	}

	products(func: foreach(ED)) @normalize {
		T as num
		to {
			i as num
			product: math(T * i)
		}
	}
}

Result

{
  "data": {
    "products": [
      {
        "product": 3
      },
      {
        "product": 4
      },
      {
        "product": 6
      },
      {
        "product": 8
      }
    ]
  }
}

Instead of

{
	products(func: eq(name, "nodeA")) @normalize {
		name
		edge {
			T as num
			to {
				i as num
				product: math(T * i)
			}
		}
	}
}

{
  "data": {
    "products": [
      {
        "product": 12
      },
      {
        "product": 9
      },
      {
        "product": 12
      },
      {
        "product": 9
      }
    ]
  }
}

Solving Queries (facets, aggregation, maths and so on) with foreach.

Solving this issue

https://github.com/dgraph-io/dgraph/issues/4160

As you can see the partial maths get broken due some internal behavior in Dgraph’s query. To not try to change the design internally. We can add foreach to solve this.

See this query

{
	Anne(func: has(rated), first: 1, offset: 0) {
		name
		rated @facets(r as rating)
		partial_sum: sum(val(r))
	}
	Brian(func: has(rated), first: 1, offset: 1) {
		name
		rated @facets(rb as rating)
		partial_sum: sum(val(rb))
	}
}

With this query, I can overcome the issue related in the report.

I gonna have this result

{
  "data": {
    "Anne": [
      {
        "name": "Anne",
        "rated": [
          {
            "rated|rating": 2
          },
          {
            "rated|rating": 5
          }
        ],
        "partial_sum": 7
      }
    ],
    "Brian": [
      {
        "name": "Brian",
        "rated": [
          {
            "rated|rating": 2
          }
        ],
        "partial_sum": 2
      }
    ]
  }
}

BUT if I use foreach, I can “simulate” this query above by doing

{
	A as var(func: has(rated))

	q(func: foreach(A)) {
		name
		rated @facets(r as rating)
		partial_sum: sum(val(r))
	}
}

This foreach query is basically creating two queries (cuz theres is two nodes “Anne” and “Brian”) internally and merging the results into one single result.