[IMPALA-4600] qgen: refactor discrepancy_searcher.py

• Type: Task
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Affects Version/s: Impala 2.7.0
• Component/s: Infrastructure
• Labels: query_generator query_generator_development
• Target Version: Product Backlog
• Assignee: Unassigned
• Reporter: Michael Brown
• Created: 06/Dec/16 22:11
• Updated: 28/Sep/17 19:50

Description

The query generator design could use improvement. One of the design principles I've heard is "an object should have 1 responsibility". That is not the case with the query generator, especially with the QueryResultComparator and QueryExecutor in disrepancy_searcher.py. They each have several responsibilities. Also, Leopard's Job() duplicates some of this work.

There should be 1 thing that can compare two Query objects' results.

There should be 1 thing that knows how to execute any AbstractStatement concurrently on both a test and reference database.

There should be 1 thing to track so-called "execution modes". Really what we want here is a sequence of setup statements, a sequence of test statements, and a sequence of teardown statements.

As it stands the above is all mingled together across the classes, make extensibility difficult.