CompositeSpecInput

A composite: up to two datasets joined on one conformed dimension, with metrics drawn from each and formulas across them, executed server-side.

input CompositeSpecInput {
  sources: [CompositeSourceInput!]!
  join_on: String!
  fields: [String!]!
  formulas: [CompositeFormulaInput!]
  mode: CompositeJoinMode
  having: [CompositeHavingInput!]
  labels: [String!]
}

Fields

CompositeSpecInput.sources ● [CompositeSourceInput!]! non-null input

CompositeSpecInput.join_on ● String! non-null scalar

The conformed dimension to join on, by its key — one every source carries (see conformed_dimensions on each dataset). Rows from the two sides meet where their values for it are equal; a value present on one side only keeps its row with the other side's fields null.

CompositeSpecInput.fields ● [String!]! non-null scalar

Output columns after the join dimension, in order: alias.field references and formula keys. A selected field left out here is still fetched (a formula may read it) but is not returned.

CompositeSpecInput.formulas ● [CompositeFormulaInput!] list input

CompositeSpecInput.mode ● CompositeJoinMode enum

Which join values to keep. Omit for ANY, the full outer join.

CompositeSpecInput.having ● [CompositeHavingInput!] list input

Keep only joined rows whose metrics pass every condition, judged before paging and the total.

CompositeSpecInput.labels ● [String!] list scalar

Dimensions to return beside the join key, as alias.field from one source. Only fields with one value per join value qualify: on contact, contact_email, contact_first_name and contact_last_name; on account, account_name. They follow the join dimension in the output.

Member Of

compositeDatasetRows query