Data Modifications under Read Committed Snapshot Isolation
Why and when UPDATE and DELETE statements acquire locks under RCSI.
Articles
Page 6 of 10
— showing articles
91 to
108
of 177
Read Committed Snapshot Isolation
The detailed behaviour of RCSI compared with locking read committed isolation, including subtle aspects to be aware of.
The Read Committed Isolation Level
The logical and physical guarantees provided by read committed isolation.
The Repeatable Read Isolation Level
The guarantees provided by the repeatable read isolation level, including some allowed behaviours you might find surprising.
Cardinality Estimation for Disjunctive (OR) Predicates
The cardinality estimation formulas for multiple OR predicates in SQL Server 2014 and later (using the 'default' cardinality estimator).
The Serializable Isolation Level
Guarantees provided by the strongest transaction isolation level and what they mean in practice.
The ACID Properties of Statements & Transactions
Concurrency considerations and the practical meaning of each component of the ACID acronym in a SQL Server context.
StarJoinInfo in SQL Server Execution Plans
Details and demos for the StarJoinInfo structures sometimes seen in SQL Server execution plans associated with an automatic performance optimization.
Cardinality Estimation for Multiple Predicates
How selectivity is computed for multiple filters in the original and default SQL Server cardinality estimation models.
Statistics on Indexed Views
Multiple reasons to prefer using the NOEXPAND hint on indexed views, even on Enterprise Edition.
Parallel Execution Plans—Branches and Threads
How SQL Server row mode parallel execution plans work.
Nested Loops Prefetching
SQL Server nested loops join internals including ordered and unordered delayed prefetching and undocumented trace flags.
Parameter Sniffing, Embedding, and the RECOMPILE Options
Internal details of SQL Server parameter sniffing, embedding (PEO), and parameterization.
Incorrect Results Caused By Adding an Index
A SQL Server incorrect results bug with merge join, partitioning, and a descending index.
Incorrect Results with Merge Join
A detailed analysis of a SQL Server optimizer bug that causes incorrect results when a descending non-unique index exists on a partitioned table.
Two Partitioning Peculiarities
Join and group by clauses may need to be written in a specific way to get the best results from the SQL Server query optimizer when using partitioning.
Aggregates and Partitioning
Improving performance for SQL Server MIN and MAX queries on partitioned tables and working around query optimizer limitations.
Working Around Missed Optimizations
SQL Server cardinality estimation bugs with partial aggregates and union all concatenation. Merge and concat union details.