Paul White explains why the IGNORE_DUP_KEY index option on a clustered index can result in much poorer performance than for a nonclustered unique index.
Aaron Bertrand (@AaronBertrand) kicks off his "Performance Myths" series, showing a "redundant" non-clustered index outperforming the clustered index with the same key.
In an effort to make leading wildcard searches sargable, Aaron Bertrand (@AaronBertrand) plays around with a trigram-type implementation in SQL Server.
Rob Farley (@rob_farley) digs deeper into seeks and actual rows read, demonstrating a case where you want a non-sargable expression in the WHERE clause.
Rob Farley (@rob_farley) looks at information exposed in query plans about residual predicates and actual rows read, showing how Plan Explorer helps identify the issue.
