SQL Server performance mistakes hide in the settings, the files and the code. This post checks four server settings, shows poor file growth and its fix, and measures two code mistakes, a function on an indexed column and a data type mismatch, in logical reads.
Spotting Definition Drift With HASHBYTES on Module Code
Did someone change a production procedure and forget to tell you? Hash the module text, compare it with a baseline, and let SQL Server point at the culprit.
SQL SERVER – Testing Database Performance with tSQLt and SQLQueryStress
Database testing is often skipped until there is a problem. Here is my approach to testing database performance with tSQLt and SQLQueryStress.
Columnstore Segment Elimination: Loading Data So Scans Skip Work
See how segment elimination skips columnstore work by checking segment metadata, measuring scans, and loading rows in a useful order.
Why ’10’ + 5 Returns 15: Mixing Strings and Numbers
The plus sign adds or glues text depending on the types on each side. I show the surprise, the error 245 it can cause, and how text keys get hurt.
Find and Replace With Regular Expressions in SSMS
A regex replace can turn forty column names into ISNULL expressions in seconds. Here is the pattern, the safety habits, and the trap of a zero default.
The Row Version Tag: Page Splits After Turning On RCSI
Turn on RCSI and every row you update gets a little bigger. On packed pages that means page splits, and I show the numbers on a small demo table.