The Heap™ – Consultancy ©®

« first day (2538 days earlier) ← previous day next day → last day (2627 days later) »

sp_BlitzErik

00:52

@JoeObbish is it an arrest warrant for Evan?

Evan Carroll

Ahh he wants all the pigs to go looking for his lost <3

Paul White

01:42

@yper-trollᵀᴹ resurrected so the original link will no longer produce an error.

If you find any more broken ones like that please custom flag the "moved to chat" comment.

→ 2 messages moved to Trash

Joe Obbish

02:26

@PaulWhite Any thoughts on something that could be added to a query that keeps the estimated execution plan mostly the same but makes the query return no results in basically 0 ms?

Something like a startup expression filter that's always pushed to the right place in the plan

sometimes the positioning of the filter leaves something to be desired...

Paul White

@JoeObbish SET NOEXEC ON;?

or any of the others that compile but don't execute the query

or any of the first_result_set procs

Depends a bit on exactly what you want to acheive and why

Joe Obbish

I want a plan in the plan cache

SET NOEXEC ON will probably work

yeah, I think that's perfect

Thanks!

4 hours later…

Joe Obbish

07:00

@PaulWhite @sp_BlitzErik Either one of you notice EstimateRowsWithoutRowGoal="6.4516e+006" in the query plan cache yet?

I'm on 2017, not sure when it was introduced

not sure why it doesn't show up if you just get an estimated plan

2 hours later…

gbn

08:41

@EvanCarroll DW shit. Real persons do OLTP. DW is for people who can wait 3 hours for a query to run

Paul White

09:07

@JoeObbish connect.microsoft.com/SQLServer/feedback/details/1008425/…

No other public information at this time, as far as I am aware.

6 hours later…

sp_BlitzErik

15:15

@JoeObbish are you on cu2? it's not in the schema yet

Joe Obbish

15:43

@PaulWhite yeah, that's what I thought it might be

@sp_BlitzErik yes

do you see mismatches between the plan cache and the estimated plan in SSMS often?

sp_BlitzErik

can't say i've compared the two much

unless i was purposely trying to get them to be different

Joe Obbish

well it's good that you haven't noticed much

sp_BlitzErik

i notice some things.

like vandalizing edits.

Joe Obbish

I already gave you a source on that one

I didn't write the rules

2 hours later…

Evan Carroll

17:46

How is everyone doing on this sacred Day of Evan.

5 hours later…

Evan Carroll

22:38

is it a defined behavior when you bitwise and bitstrings of unequal length

ah nvm, doesn't even allow it

ERROR: cannot AND bit strings of different sizes

Evan Carroll

23:20

A: Lookup performance of Numeric vs String

Hate to be captain obvious on this one, but they even provide a function to do this that stores the intergers as bigint.. CREATE SCHEMA insta5; CREATE SEQUENCE insta5.table_id_seq; CREATE OR REPLACE FUNCTION insta5.next_id(OUT result bigint) AS $$ DECLARE our_epoch bigint := 1314220021721; ...