0008 datatable driven parallel execution

8. Datatable driven parallel execution

This ADR has been recorded retrospectively.

Status

Accepted

Context

During parallel execution, each specification is executed in a separate runner. Specifications which have a datatable under then are also run by a single runner. If there are many rows in a specification, it takes the runner quite sometime to finish execution.

Decision

This can be optimized, and run each row of a datable driven specification in a separate runner.
A new spec is created for each row in a table and will be executed in parallel. Scenarios which do not use datatable rows are to be executed once. All such scenarios will be added to the first spec created.
Reports should show a single specification, irrespective of it being datatable driven. All results need to be merged to enable this.
Before and after spec should be executed once per specification created. This is a change in behaviour where before, they were executed only once as all table driven scenarios used to run in a single spec.

Consequences

__