Skip to content

Decision Tables ​

A decision table is a step of type decision_table. Each row is a rule: input cells on the left, output values on the right. It is often shorter and easier to review than a chain of decision steps.

json
{
  "id": "discount",
  "name": "Discount",
  "type": "decision_table",
  "hit_policy": "first",
  "inputs": ["customer.tier", "order.amount"],
  "outputs": ["discount", "reason"],
  "rules": [
    { "when": ["gold", ">= 1000"], "then": [0.15, "Gold, large order"] },
    { "when": [["silver", "bronze"], "100..999"], "then": [0.05, "= customer.tier + \" tier\""] }
  ],
  "default": [0, "No discount"],
  "next_step": "price"
}
  • inputs are expressions, one per column.
  • outputs are variable names. Later steps read them as $discount and $reason.
  • Rows are checked top to bottom.

The same table in YAML:

yaml
discount:
  id: discount
  name: Discount
  type: decision_table
  inputs: [customer.tier, order.amount]
  outputs: [discount, reason]
  rules:
    - when: [gold, ">= 1000"]
      then: [0.15, "Gold, large order"]
    - when: [[silver, bronze], "100..999"]
      then: [0.05, "= customer.tier + \" tier\""]
  default: [0, "No discount"]
  next_step: price

Input cells ​

CellMatches when
"*", "-", "" or nullalways
"gold", 42, truethe input equals the value
["silver", "bronze"]the input is one of the values
">= 1000", "< 18", "!= \"x\""the comparison holds
"in [...]", "not in [...]"membership
"18..65"the number is in the range, ends included
"= <expression>"the expression is true (it can use any field)

A plain string is compared as text. To compare with a string that starts with an operator or with in , write "== \"in stock\"".

Output cells ​

A literal (0.15, "Gold", null, an array or an object), or "= <expression>" to compute it, such as "= order.amount * 0.1".

Hit policies ​

hit_policyResult
first (default)The first matching row sets the outputs. With no match, default is used. Without a default, execution fails instead of going on with missing values.
collectEvery matching row, in order. Each output becomes an array, or a single value when aggregate is sum, count, min or max.

With collect and no match, outputs are [], sum and count are 0, and min and max are null.

Adding up risk points from every row that matches:

json
{
  "id": "risk",
  "name": "Risk points",
  "type": "decision_table",
  "hit_policy": "collect",
  "aggregate": "sum",
  "inputs": ["order.amount"],
  "outputs": ["points"],
  "rules": [
    { "when": ["> 100"], "then": [10] },
    { "when": ["> 500"], "then": [20] }
  ],
  "next_step": "decide"
}

For an amount of 600, $points is 30.

Studio also has a table view for editing rules as a table. See Studio.

Decision table steps cannot be compiled to the binary .ordo format yet.

Released under the MIT License.