We have applied query object patterns to isolate query objects from the repository, which maintains query logic in one single place. We created a simple query object, then invented a repository to illustrate how query objects could be accommodated in the repositories. We then moved from queries to commands to perform real transactions using a similar pattern. Finally, we worked with expression trees to simplify and configure the conditions in a query execution.
Let's discuss how we could handle failure scenarios using transactions, how the transaction could be accessed between contexts, and also between different data access technologies:
- Is there any complex system that performs more than one data manipulation activity, then handles failures to maintain atomicity?
- Transactions allow us to handle atomicity in failure...
