Read operator cost percentages as estimates, then use actual rows, elapsed time, IO, and repeated work to find the real bottleneck.
Rolling Date Ranges Forward Without Hard-Coded Years
Replace hard-coded years with indexed date ranges for current periods, full months, year-to-date comparisons, and any requested year.
Why One Query Has Two Cached Plans: SET Options and plan_attributes
Investigate two cached plans with plan_attributes, decode differing SET options, and match application settings before diagnosing skew.
Sharing an Execution Plan Without Exposing Sensitive Data
Protect sensitive data when sharing an execution plan by inspecting XML values, redacting names and literals, and checking the file again.
Trigger Overhead: Measuring the Hidden Cost of DML Triggers
Measure trigger overhead with controlled inserts, execution plans, and cached trigger statistics before moving slow work into a queue.
Filtering on Dates Without Losing the Index
Filtering on dates without losing an index means using searchable ranges; learn half-open boundaries, date and datetime differences, and plan checks.
Reading the Warnings in an Execution Plan
Warnings in an execution plan can reveal conversions, missing statistics, and spills; read each warning in workload context.