Given the same data, schema, the same query against two different table configurations should always produce the same result. Generating table configurations over a schema and checking this invariant for generated random data would help to find bugs in operators and index implementations.
Given the same data, schema, the same query against two different table configurations should always produce the same result. Generating table configurations over a schema and checking this invariant for generated random data would help to find bugs in operators and index implementations.