I'm designing Sticky notes Management Application (CRUD), where StickyNote has 3 attributes, heading, summary and task list. The task has 2 attributes - heading and summary.
A sticky note can have upto 1000 tasks.
The class would look like -
public class StickyNote {
String id;
String heading;
String summary;
Set<Task> tasks;
}
When it comes using Database , to me a NoSql/Document structure seems more close to OOPS.Thus having one table representing StickyNote, with Task as a field/column.
Pros of this approach -
Deletion - As multiple sticky notes can have same tasks, thus deletion of task could be straightforward, by providing task heading and sticky notes id.
No need of introducing a second table for tasks. Hence either avoiding duplicate records or foreign key constraint check while deleting.
Update - a task would be updated either using task heading or task summary. Thus iterating over the task list, and as max tasks could be 1000, performance won't be an issue.
Cons -
- Using id for deleting task is straightforward and performance efficient, rather than using iterator.
Questions-
- Would introducing Task table would be performance efficient?
- Would relational approach be some how better than non relational one?
- If I consider to scale up application, using million users in concurrent environment, introducing a Task id would be mandatory as iteration would then degrade performance, which essentially would lead to add a separate Task Table. Hence in that case both nosql and sql approach would be same?
