The uniquifier is a hidden 4-byte value that SQL Server adds to duplicate keys in a non-unique clustered index. This post measures its real cost row by row on SQL Server 2025 and shows it on the page. It also compares it with a unique key you choose yourself.
Scan Count Zero in STATISTICS IO: What the Number Means
Scan count zero in STATISTICS IO is not an error and not a free read. This post runs nine small queries and shows what 0, 1, 2 and 5 mean, why a unique lookup counts as zero, and why logical reads are the number to compare.
Alerting on Long-Running Queries With a SQL Agent Job
Alert on long-running queries with a five-minute SQL Agent job, a deduplicated log, and one Database Mail notice per new request.
Oldest Statistics in SQL Server: Find the Stale Ones First
To find the oldest statistics in SQL Server, read the last update time and the number of changes since. This post builds a small demo, lists every statistic with its age and percent changed, and shows why a high change count on its own is not a reason to update by hand.
Performance Regression After a SQL Server Upgrade: First Aid
A performance regression after an upgrade hurts most when users are waiting. This post shows a first-aid switch that brings back the old row estimates without lowering the compatibility level, and proves with a demo what changes and what stays. It also covers the undo, the per-query options and the steps that find the real cause.
Partition Switch in SQL Server: Move a Million Rows in a Moment
A partition switch moves a whole partition from one table to another without copying a single row. This post builds a partitioned table with a million rows, switches the January partition out and back in, and shows that only metadata changes. It lists the five errors that stop a switch, with the real message numbers, and ends with the sliding window pattern.
BETWEEN vs IN in SQL Server: Which One Reads Less?
BETWEEN vs IN looks like a style choice, and it is not. This post builds a demo table and measures BETWEEN, the two operators and an IN list with and without an index. It shows that BETWEEN and the operators share one plan, that IN seeks once for every value and that sparse values change the question.







