Skip to content

Data objects and data stores

Two drawing elements that describe what a process works with, without doing anything in it: a data object (a document or a piece of information that travels) and a data store (a place where information is kept between cases).

Both are in the canvas toolbar. A data object takes a label, and its type is set from its context menu — Empty, Input or Output, which only changes the icon drawn.

They do not carry data

This is the rule worth remembering: a data object or a data store does not store, move or create anything. The engine ignores them when a case runs.

Where the data actually lives:

What the diagram shows Where it lives in the product
An input document Input documents attached to a step
A generated file Output documents
A record, a list, a reference table Case Tables or a process variable
A value read from another system A database connection used by a control or a trigger

A diagram that draws data flowing from a data store into a task does not read anything when the case runs; it only tells the reader that it should.

When they earn their place

  • Explaining a process to someone outside the team. A data object labelled Invoice next to the task that receives it saves a paragraph.
  • Mapping an existing process. Drawing the documents and the registers makes the gaps visible before you build anything.
  • Reviewing a design. If the diagram shows a data store that no task writes to, the process probably has a missing step.

Avoid using them as if they were the implementation. The diagram is checked against the engine by reading the task assignments and the variables, not by following the data objects.