Уровень 0 · материалов: 3
В кластер входят документы, посвященные техническим аспектам реализации и оптимизации функций MS SQL Server, и не входят документы по общему администрированию или другим СУБД.
Общие признаки: производительность SQL запросов, технические детали работы СУБД, оптимизация работы с данными
Группа выше: Индексы и настройка СУБД
Смысл: The main idea is that while SELECT COUNT(*) is the standard way to count rows, it is inefficient for large tables; using system metadata is significantly faster but comes with rare risks of inaccuracy that can impact the query optimizer.
The text compares different SQL Server row-counting methods, proving that querying system metadata is vastly faster than COUNT(*) for large tables, despite potential synchronization issues.
Смысл: The main idea is to provide a data-driven comparison of primary key strategies in MS SQL Server, proving that sequential GUIDs generated on the client side offer the best balance of functionality and performance.
Sequential GUIDs generated on the client side are the most performant choice when a GUID is required as a primary key in MS SQL Server.
Смысл: The main idea is to clarify the technical difference between ROWS and RANGE frames in SQL window functions, emphasizing that ROWS operates on physical row counts while RANGE operates on logical value ranges.
A technical explanation of MS SQL Server window function frames, detailing how ROWS and RANGE differ in handling non-unique sorting values.