Вот цитата из выводов статьи:
We have also shown that relational database systems produce large
estimation errors that quickly grow as the number of joins increases,
and that these errors are usually the reason for bad plans. In contrast to cardinality estimation, the contribution of the cost model to
the overall query performance is limited.
Going forward, we see two main routes for improving the plan
quality in heavily-indexed settings. First, database systems can incorporate more advanced estimation algorithms that have been proposed in the literature. The second route would be to increase the
interaction between the runtime and the query optimizer. We leave
the evaluation of both approaches for future work.
Инновационные предложения состоят в "Напишите более лучшие оптимизаторы" и "Собирайте более лучшую статистику". Спасибо, пойдем работать и делать! Без этих советов мы бы сами об этом никак не догадались 🙂
We have also shown that relational database systems produce large
estimation errors that quickly grow as the number of joins increases,
and that these errors are usually the reason for bad plans. In contrast to cardinality estimation, the contribution of the cost model to
the overall query performance is limited.
Going forward, we see two main routes for improving the plan
quality in heavily-indexed settings. First, database systems can incorporate more advanced estimation algorithms that have been proposed in the literature. The second route would be to increase the
interaction between the runtime and the query optimizer. We leave
the evaluation of both approaches for future work.
Инновационные предложения состоят в "Напишите более лучшие оптимизаторы" и "Собирайте более лучшую статистику". Спасибо, пойдем работать и делать! Без этих советов мы бы сами об этом никак не догадались 🙂