The 10-score answer would mentioned that T-SQL language in terms of line-by-line execution is effectively an interpreted language and for the technical reasons of how the execution plans are being parsed/compiled/executed (add details) or otherwise, the...

Explain why doing 4 variable assignments in one line runs in half the time as doing 4 variable assignments in 4 lines in T-SQL....

Shared allocation and interleaving can cause indexes to be rebuilt in a fragmented state. According to a link provided in another answer to this question, in SQL Server 2005, indexes with less than 24 leaf level pages can get shared allocation for th...

Which of the following are correct for "Foreign Keys"? 1. Enable the data to store foreign languages that require Unicode data types 2. Should generally have non-clustered indexes 3. Should never be nullable 4. Should be included in the cluste...

My guess is that databases on mobile devices (assuming SQL Server Compact is like SQLite) are largely single-user, and opening a file rather than connecting to a server, and the DBMS code being compiled into the application instead of the application co...

Some poorly designed semi-structured schema systems can make very good use of sparse columns. Normalized semi-structured schemas are a pain to ETL into though. For normalized semi-structured schema systems that want to avoid using sql_variant, spars...

What's faster depends on how selective the report is. With High-selectivity reports, the nested loop join with an index seek on its right side is fastest. Vs the cost of a scan, the inefficiency of random-access I/O is justified if the # of iteratio...