Is there no way you could anonymise the query plan and tables before sending them up?

If the parameter can be linked back to the statistics, it's still going to be related to the row count.

I'm assuming in an ideal scenario you want correct...

Given that memory grants are based on row counts, and row counts are assessed to actually build the execution plan itself, it's pretty safe to assume these are cached.

So are you guys saying that in effect the queries will take longer to run and consume more CPU if they are parallel vs non-parallel? (For instances of small queries)

Sean Lange (12/15/2016)

---

george.p (12/15/2016)

---

I missed the title so I assumed it meant Scalar functions 🙁

You're one of those people that puts all of the details in the title of an...

I missed the title so I assumed it meant Scalar functions 🙁

You're one of those people that puts all of the details in the title of an email aren't you!...

This code is very difficult to understand - there are no supporting comments and a lot of variables named as single letters.

Some documentation of how it works would be useful!

Based on this excerpt from MSDN the SP_Recompile will just clear the plan cache for the associated object.

The sp_recompile system stored procedure forces a recompile of a stored procedure the...

Ah right I see - I assume it's two things then:

1. If the log file is filled auto growth would slow things down

2. Minimal logging physically writes less to the...

Why would setting the recovery mode to simple make bulk inserts faster?