Loading partitions in parallel uses separate staging tables, controlled workers, and partition switching to publish data without chaotic locks.
Getting Through a Technical Book You Actually Finish
Read a technical book with a practical question, small SQL experiments, and reusable notes that turn each useful chapter into working knowledge.
Plan Cache Key Lookups: Finding the Worst Ones With XQuery
Not every key lookup needs fixing. Rank statements by reads, find the lookups with XQuery, then test the smallest index change on one call.
MAXDOP 1 in SQL Server: SQL in Sixty Seconds #008
MAXDOP 1 requests serial execution for one SQL Server query. I measured four variants of one query on SQL Server 2025, and the serial plan was the slowest in elapsed time with about the same CPU. The post shows the settings behind the hint and how to read the degree of parallelism a query actually used.
Computed Column UDFs: Why Every Query on the Table Goes Serial
A computed column that calls a scalar function can block parallel plans for queries that never read it. Here is how to spot it and test a fix.
Table Variables and Transactions in SQL Server: SQL in Sixty Seconds #007
Table variables and transactions do not mix the way temporary tables do, because a ROLLBACK leaves the rows in a table variable. The tests here cover a full rollback, a savepoint, a failed statement and a logging pattern, all on SQL Server 2025.
Why a Second Log File Does Not Make Anything Faster
Adding a second log file feels like a speed fix, but it is only a spare room. Here is how to watch it work and how to retire it.