Comments for OraStory

Comment on SQL Patch IV – Why is the SQL Patch applied but not effective? by Dynamic Sampling – 2 | Oracle Scratchpad

Dynamic Sampling – 2 | Oracle Scratchpad — Thu, 23 May 2013 12:47:26 +0000

[…] part 4 […]

Comment on SQL Patch III / Parallel Baselines by Dynamic Sampling – 2 | Oracle Scratchpad

Dynamic Sampling – 2 | Oracle Scratchpad — Thu, 23 May 2013 12:47:24 +0000

[…] part 3 […]

Comment on SQL Patch II by Dynamic Sampling – 2 | Oracle Scratchpad

Dynamic Sampling – 2 | Oracle Scratchpad — Thu, 23 May 2013 12:47:22 +0000

[…] part 2 […]

Comment on SQL Patch I by Dynamic Sampling – 2 | Oracle Scratchpad

Dynamic Sampling – 2 | Oracle Scratchpad — Thu, 23 May 2013 12:47:19 +0000

[…] part 1 […]

Comment on ORs, IN lists and LNNVL by Parse Time | Oracle Scratchpad

Parse Time | Oracle Scratchpad — Mon, 13 May 2013 18:59:45 +0000

[…] Brooks published a note recently about some very nasty SQL – originally thinking that it was displaying a run-time problem due to the extreme number of […]

Comment on ORs, IN lists and LNNVL by Jonathan Lewis

Jonathan Lewis — Thu, 09 May 2013 12:43:41 +0000

Dominic,

I built an approximate model based on reverse engineering your execution plan before I saw the offer. I’ll try to find time to wrap it in a little text and post it later on tonight.

Comment on ORs, IN lists and LNNVL by Dom Brooks

Dom Brooks — Wed, 08 May 2013 20:28:07 +0000

This was the problem I had with my initial “demo” – I struggled to create a test case away from the application that had the same impact (now known to be parsing).

If you want to take it further, would it help if I emailed the original query or anything else?

Comment on ORs, IN lists and LNNVL by Jonathan Lewis

Jonathan Lewis — Wed, 08 May 2013 20:00:24 +0000

Dominic,

Thanks for the follow-up.

I was struggling to figure out how the execution could take so long – now I’m struggling to figure out how the optimisation could take so long ;(

I may have to write a program to create a dataset and (more importantly) a query that’s insane enough to emulate your monster.

Comment on Thoughts on TM_DELTA_TIME by Steve Howard

Steve Howard — Wed, 08 May 2013 13:41:17 +0000

If I am doing the math correctly, it looks like there 985 seconds in your first session record in ASH at 15:05. This is about 16 minutes. Your previous record was 2:45PM, or about 20 minutes prior. I am wondering if you didn’t do something else in your session at about 2:49PM, and then nothing for 16 minutes. Could there be a bug where the time is recorded since the previous measurement? I know that defeats the entire purpose of tracking “active” sessions, but the math at least makes me ask the question.

Comment on ORs, IN lists and LNNVL by Dom Brooks

Dom Brooks — Wed, 08 May 2013 08:47:52 +0000

Jonathan,

Spot on.

NAME                                                             VALUE
---------------------------------------------------------------- -----
parse time cpu                                                   28402 
parse time elapsed                                               28475 

Elapsed: 00:00:00.016
---
run query
---

Elapsed: 00:01:57.938

NAME                                                             VALUE
---------------------------------------------------------------- -----
parse time cpu                                                   40022 
parse time elapsed                                               40102 

Elapsed: 00:00:00.000

Sloppy initial analysis…

Thanks.

Cheers,
Dominic