Type: Database refactoring

A special case of Split Table


As one step of Extract Bounded Context you’ve found a monolithic data model. (Often used to store a monolithic domain model.) In this data model you have identified a table that has become too big. The decision has been made to separate the parts for the newly extracted bounded context.

Schema Update Mechanics

  • Create an empty new table in the database scheme of the carved-out context.
  • Copy to-be-duplicated columns from old to new table.
  • Copy to-be-moved columns from old to new table by repeatedly applying Move Column.
  • Introduce synchronizing triggers on both the monolithic and the carved-out table. They must be invoked by any change to the data in the tables. Beware of the triggers triggering each other.
  • Add a drop date to the triggers.

Data-Migration Mechanics

  • Copy the data from the duplicated columns to the new table. Potentially not all rows existing in the monolith should also be present in the carved-out bounded context.
  • Copy the data from the moved columns to the new table with Move Data.

Access Program Mechanics

  • Analyze the accessing program code thoroughly and update it accordingly during the transition period.
  • If applicable extract a specialized repository for the carved-out bounded context.