TomThomson (5/15/2014)

---

sknox (5/15/2014)

---

Setting aside the trace flag issue (I don't blame the question writer for Microsoft's inability to write proper documentation) the only problem I have with this question is the use of the term "best".

To paraphrase an old quote about lawyers, ask five SQL Server professionals the best way to write a simple SELECT statement, and they'll give you six(teen) different answers.

No, I wont believe 16 and certainly not 6. 5 SQL Server professionals will never all agree on a compromise solution, so the correct answer is not 32, and zero SQL Server professionals won't offer an answer at all, so it isn't 31 either. So it must be 30, since any two disjoint nonempty sets of SQL professionals will provide two different answers, and no proper superset of some (non-empty) original set of SQL Server professional will provide the same answer as the original set.

Of course this only applies to small sets, as can clearly be seen by considering the subsets of a set or 20 SQL Server professionals, especially if one places a limit on the amount of irrelevant twaddle included in the answer, since there are many SQL Server questions for which it just isn't possible to imagine 1048754 different answers, but it's safe for sets with fewer than 10 members since SQL Server Profesionals can devise 510 answers to any question about SQL Server.