This Policy Based Management Quiz asks which evaluation mode can stop a bad change before it is saved. The answer is On change: prevent. A tested T-SQL script blocks a badly named table and shows how few facets support that mode.
SET NOEXEC and PARSEONLY: Checking a Script Without Running It
Use PARSEONLY and NOEXEC to inspect scripts without executing them, understand deferred references, and check existing module dependencies.
Common Table Expressions for Readability
Common table expressions for readability name query steps and support recursion, but they do not promise cached intermediate results.
SELECT * in Production Code: The Hidden Costs
See the real costs of SELECT * in production code: lost covering indexes, lookups, stale view metadata, broken inserts and wider results.
Inclusive End Dates: Why BETWEEN Misses the Last Day
Inclusive end dates describe whole calendar days, while BETWEEN compares the exact supplied endpoints. A date endpoint becomes midnight when compared with timestamp rows. A guarded disposable fixture makes the boundary conditions explicit.
Keyset Paging: Continue After a Date and ID Cursor
Keyset paging continues after a saved ordering key instead of counting past an offset. A date and unique ID together keep equal-date rows in the sequence. A guarded disposable fixture makes the boundary conditions explicit.
Storing Percentages: 0.15 or 15, and How to Tell Them Apart
Storing percentages starts with a declared unit. Fractional rates and percentage points require different conversions, even when both pass a range check. Use named columns, explicit boundaries and a guarded disposable fixture.