Piet de Visser

Probably Correct.

But also: maybe want to keep this a guarded secret to avoid a grey-wave ?

??

The Green Barrel will do.

(in an earlier life, my Boss would proudly show that green thing... )

Grinning...

but everyone want to hear + see AI-stuff..
we have to at least pretend some AI.

it is abt mandatory #buzzwordcompliance ...
Even by criticizing the #buzzwordcompliance, you can increase your count of hashtag-hits with #buzzwordcompliance...

How do you apply "Green" practices in your work ?

Desired answer: we buy Expen$ive green paint...
(bigger budget is more Status for Boss, more commi$$ion for $alesrep)

My Answer: make systems More Efficient, use less resources, consume less electricity....
(not always what they wants to hear...)

There is an Old Proverb...

To Find Typos... Publish your blog.

And Thank You !

Also : Congrats with you recent career-moves.

I'm now running an image provided by an Oracle-VP...
(and using microdnf to tweak it a bit...)

hm.
today we still laugh at this....

hmmm
if you didnt manage to "hand over" to some minion, it may come back to haunt you.

The uninformed will blame you for what they dont understand...

Oracle23ai and python - How to fix (avoid) a MERGE statement. TL;DR: Still using Oracle and python, and loving the combination. But some Annoying High-Frequency MERGE statements from several python prog...

Ancient Wisdom to make a program faster: Dont Execute it.

How to skip "unnecessary" code, and thus avoid RoundTrips...

simpleoracledba.blogspot.com/2025/08/orac...

#Oracle
#python

Let me add... any teaching that uses Pascal, or a similar language to encourage "Structured Programming" is probably Good Stuff.

@pougorg.bsky.social is in a month.
I got my "concept" for prez accepted, but,...
They turned it into a "workshop".

Been thinking about it.
I'll try to make it interesting, with "audience participation" (BYODatabase!).

Image of me thinking...

FP Brooks, Mythical Man Month...

I'll also mention K+R.
I never was a good C-programmer, but a bit of (pro*)C-practice helped develop useful (conceptual) programming-skills, Still Valid.

Anwenderkonferenz

Das Programm der DOAG 2025 Konferenz + Ausstellung ist online! Aus den zahlreichen Einreichungen hat unser Programmkomitee eine hochklassige Agenda mit rund 300 Sessions zusammengestellt. Jetzt im Programm stöbern und Early-Bird-Tickets sichern! scomp.ly/Aw8gZoL #DOAG2025

Final Thought: as Epoch-day approaches, I can just see some AI-resultset kicking this up by 2037..

Hi There!

(we did Y2K, now you can fix Epoch... )

And the angle of Timezone..: For my epoch-functions and macros, I thought I wanted UTC (I think epoch is defined as seconds since 1-1-1970-00:00:00 At UTC.. ?)

Any thoughts ?
(do we need to discuss this over beer in at some confrnce...)

re-1: constant-value can be achieved with macro, with CTE-materialize, with sql_transpiler, .. probably more...

re-2: dynamic (geeky? normally would not want...) can be achieved from just a function, see my earliest attempts.

but "dynamic" might expose some inner-workings.. hehehe. curious.

There are many dimensions to this.

One Angle:

1- to follow sysdate + systmstmp, you want a Fnctn that returns "constant" during exec of the SQL (call that atomic)
2 - to measure "inside" a DML (or Select), you want a Fnctn to return point in time at exact time of call. quite geeky/curious.

hmm, I was probably mistaken..
Cant reproduce.

When I test now (script or manual-typing)
set trnsplr=on : all f_epoch results are equal, and match systmstmp. Nice!
set trsplr=off (bck to dflt) has immediate effect: all f_epoch values vary. even with identical stmnts.

As Expected, I think.

Funny but serious...

Many systems HANG !

Checking... Lock-Wait, Find blocker-session...

Someone had (justly) updated parent-data via a SQL-tool, session still open... (forgotten to hit commit-button...)

Grinning at "lifetime learning" ...
And thinking: "the other system has autocommit...

Thx. I wondered about that already.
Had also expected the need for explicit CAST 2 date ?
I'll look into that.
The function f_epoch2 came from some quick trial+err.
And using sysdate (in the function version) to calculate "whole-seconds-to-midnight" was notably faster than using 2 x systimestamp.

"normal", as in: Each call to the function returns a slightly diff epochs (visible in fractional seconds).

But that behaviour only returns After the statement is re-parsed with transpiler reset to off.
The parse with identical epochs can survive the alter-session if you dont edit/parse the stmnt.

And a sneaky by-effect of parsing: if you turn the transpiler OFF, the behaviour returns to normal, but NOT (yet) for the SOFT-Parse stmnts. You have to Force a Hardparse for the behaviour to return completely.

This Database will Always Amaze...

Brilliant.
Why didnt I know this...
(I know, I know.. RTFM... and then some...)

Messing about with "code"...
Can just say that the art of #defensiveprogramming needs too be re-visited...

Nobody seems to check for (input-)errors much, and that is potentially a (security)risk.
I guess a lot of "incidents" are either not-noticed, or kept silent.

en.wikipedia.org/wiki/Defensi...

Le Mot (@WordleFR) #1298 5/6

⬛🟨⬛🟨🟨 good starter word eh..
🟨🟨🟨⬛🟨 Hmm.
🟩🟨🟨🟨⬛ now a lot of clues..
🟩🟩🟩⬛🟩
🟩🟩🟩🟩🟩

wordle.louan.me

on 3rd.. Wow

Storage...
I bet your storage engineer keeps saying something like:
"my io-system is not very busy, really"...
(if he, always a he, as some monitoring-capability at all...)

And software-managed IO systems may well suffer from some scheduling-effect, like "off-cpu".

Happy Puzzling !

Grinning right at you...

Every now and then we do that: run iostat, vmstat or sar and scrape the data into a worksheet, make a quick-graph of something... and Spot the Anomaly.

So many fancy monitoring-tools have tried to replace or improve.
But the quick DIY-Excel is still as good as 30yrs ago...

and yes.. it was about (eliminating) RoundTrips...