« first day (4348 days earlier)      last day (522 days later) » 

12:55 AM
@J.D. defensive programming, in case the INSERT had failed ;)
 
1:40 AM
@ypercubeᵀᴹ 🤔, OUTPUT clause instead?...then you get the atomicity of a single statement.
 
 
1 hour later…
2:50 AM
@J.D. A while back, we wanted to suppress the EF code from doing INSERTs it shouldn't have been doing (excessively verbose and unneeded logging of data we didn't need) with an INSTEAD OF trigger which only allowed certain inserts to go through. Needless to say, EF doesn't like that, so that approach had to be abandoned.
 
 
1 hour later…
4:02 AM
EF should be abandoned
4
Seriously, need a good c# data layer... Not whatever EF has become
 
4:51 AM
@CadeRoux Yea, been there done that too. Triggers with EF Core can get a little dicey. Need to make sure the right settings are applied in the trigger and that the row counts returned are what EF expects etc.
 
 
2 hours later…
7:13 AM
Morning
 
Morning
 
Morning
What do I do if my iPhone tells me: Your Apple-ID is not active.
 
7:44 AM
Activate it
@SeanGallardy @ErikDarling It wasn't 100% snark. Explaining needed some SSMS screenshots and better code formatting than is available in comments. Besides, explaining why the OP didn't understand belongs in a chat environment anyway. I've done all that now. Time will tell if the angry hacker from Stack Overflow is enlightened by it all or not.
 
Good morning Heap. Where's the best (least bad?) place these days to report a bug in SQL Server 2022 Synapse Link?
 
8:05 AM
Is that SQL Server bug or synapse bug?
 
There's no good place to report bugs
Even if you have a support agreement
Maybe make a viral TikTok about it
 
8:23 AM
@PaulWhite It used to work. Until this morning a 6am. To continue using iMessage on this iPhone, verify your password (original Apple iPhone message)
Verified my password with existing Apple-iD and boom...
 
Sounds like phishing. Maybe it's Elon trying to take down Apple
 
@Zikato ....or worse. I have been pawned.
 
Maybe you just need to pay the 30% activation tax
 
@PaulWhite Yup. After searching for 2 hours found on the support site a page that allows being called back at 11.45am
Pawned once - Gravatar
 
We've had to get in touch with them a couple of times about Apple IDs. Was never anything bad, just one of those things.
 
8:29 AM
Glad to hear that.
 
I use mine so infrequently it's a wonder I even remember the password
 
I dread being hacked, b/c all passwords on keyring, etc.
 
I hear GingerCoin solves all these problems
 
@Zikato Synapse sending code that SQL Server can't run. It's a collation problem. Plenty of error to go round.
 
aka.ms/synapsefeedback
 
8:39 AM
@PaulWhite Ooooh! I could do it in an interpretive dance hip hop stylie. Here, hold my beer. :-)
2
 
8:57 AM
@Ginger How do some of the questions differ from standard SO questions?
> What are the advantages/disadvantages of implementing strings as character arrays?
> What is a generational garbage collector and what are its benefits?
 
9:17 AM
@Zikato 👍
 
but screaming into the void might get more traction
 
 
2 hours later…
11:12 AM
@ErikDarling Weird because the multiple RETURN statements restriction was only added in CU5
If you ever get chance to repro, I'd be somewhat interested to hear how this goes:
CREATE OR ALTER FUNCTION dbo.F
(
    @a bit = 0,
    @b bit = 0
)
RETURNS varchar(5)
WITH SCHEMABINDING
AS
BEGIN
    RETURN
        CASE
            WHEN @b = 0 AND @a = 0 THEN 'S'
            WHEN @b = 0 AND @a = 1 THEN 'A'
            WHEN @b = 1 AND @a = 0 THEN 'B'
            WHEN @b = 1 AND @a = 1 THEN 'AB'
            ELSE ''
        END;
END;
That ought to be optimally inlineable (no mess of nested loop joins)
 
11:30 AM
Quite a big earthquake swarm around NZ's supervolcano this evening.
Probably nothing.
Lake Taupō, in the centre of New Zealand's North Island, is the caldera of the Taupō Volcano, a large rhyolitic supervolcano. This huge volcano has produced two of the world's most violent eruptions in geologically recent times. The volcano is in the Taupō Volcanic Zone, a region of rift volcanic activity that extends from Ruapehu in the south, through the Taupō and Rotorua districts, to Whakaari/White Island, in the Bay of Plenty. Taupō began erupting about 300,000 years ago. The main eruptions that still affect the surrounding landscape are the dacitic Mount Tauhara eruption 65,000 years ago...
The Oruanui eruption (also known at Kawakawa eruption or Kawakawa/Oruanui event) of New Zealand's Taupō Volcano was the world's most recent supereruption. == Eruption == With a Volcanic Explosivity Index of 8, it is one of the largest eruptions ever to occur in New Zealand. It occurred about 26,500 years ago in the Late Pleistocene and generated approximately 430 km3 (100 cu mi) of pyroclastic fall deposits, 320 km3 (77 cu mi) of pyroclastic density current (PDC) deposits (mostly ignimbrite) and 420 km3 (100 cu mi) of primary intracaldera material, equivalent to 530 km3 (130 cu mi) of ma...
The Hatepe eruption, named for the Hatepe Plinian pumice tephra layer, sometimes referred to as the Taupo eruption or Horomatangi Reef Unit Y eruption, is dated to 232 CE ± 10 and was Taupo Volcano's most recent major eruption. It is thought to be New Zealand's largest eruption within the last 20,000 years. The eruption ejected some 120 km3 (29 cu mi) of material, of which just over 30 km3 (7.2 cu mi) was ejected in a few minutes. This makes it one of the largest eruptions in the last 5,000 years, comparable to the Minoan eruption in the 2nd millennium BCE, the 946 eruption of Paektu Mountain,...
 
12:26 PM
@PaulWhite I apologize for the fuzzy details. It was a short discussion on a short call. They seemed to consider the matter closed with the change and wanted to move on to important topics like “what does this graph we made from a query no one here wrote mean?”
 
Invoice them extra hard for me
 
@PaulWhite if you need to evacuate you can stay in the Darling Data executive penthouse
 
great thx
I look forward to being inlined
 
Inline Bling
 
who put the L in Bing
 
12:32 PM
Some Canadian rapper
 
1:06 PM
@PaulWhite I will not stand for this bashing of nested loop joins!
 
Nothing wrong with a nested loop join
Froid goes a bit nuts though
In this particular case, one case expression is superior to six NLJs
they're all applys
not that it matters
at the risk of taking your remark seriously
 
@PaulWhite I am being facetious
Although if that isn't hitting disk too hard the difference might not be huge - depends on how important microseconds and CPU usage are to you.
 
1:23 PM
Pretty rare for constant scans to hit disk
It also depends how often it's called
In Erik's case, something was causing the inlined version to be 100x slower
But this is all new to me so I might be wrong
 
@PaulWhite can i interest you in 75% off SQL Server training?
 
TAKE MY MONEY
 
You’ll really like the 25% that’s correct
Hey, it finally hit 100!
100
A: Why is my query suddenly slower than it was yesterday?

Erik DarlingDear [your name here]! Oh no, I'm sorry to hear that! Let's start with some basics to get you fixed up in a jiffy. The thing you're running into is called Parameter Sniffing It's a way out wiggy weird problem. The name rolls right off the tongue. Like the German word for squirrel. And it's us...

Only took like five years
 
@PaulWhite There are always instances where that will be the bottleneck, but I wonder if it was also forcing things single threaded - but yes, all speculation without seeing things in detail
 
@ErikDarling Let me just move that to meta
@bbaird Inlined scalar functions don't disallow parallelism
 
1:30 PM
@PaulWhite wasn't this preventing it in-lining? Or am I misremembering the issue?
 
@bbaird The one that couldn't be inlined was the faster one
I bit my tongue the hardest I've ever done in my life this morning while eating a hot cross bun
 
@PaulWhite ah, ok. interesting - I am on a different caffeine schedule over here as you are aware
 
It's still sore now
 
Ow.
 
Yeah, probably payback for the mirror incident yesterday
I thought I noticed an extra pin in the doll today
 
1:32 PM
@PaulWhite I’ll move you to Melbourne
 
whoa man way to overreact
I wonder if I community wiki'd it, would you lose all the rep
 
Only one way to find out
 
actually, MELB isn't bad as deadly Aussie cities go
So what did you earn for that? A gold Great Question, Brent badge?
 
I think it just keeps you from getting more rep. Not that Erik can't afford it
 
Suspension takes rep down to 1
P sure threatening to move someone to Melbourne is a CoC violation
 
1:36 PM
@PaulWhite I got the “used to really care” badge
 
🤣
 
@PaulWhite Melbourne is a warning for first offense - Adelaide is instaban
 
now do Cairns
the nuclear option
 
naming things is hard
 
1:38 PM
> See my query please
1
Q: Why do Sql Server In operator behaves differently and return nothing?

tekfI am using sql server 2017. I am trying to use the in operator, but it behaves me so differently. What I am trying to do is I created a cte and split the comma separated string into array and use this cte in my end query through in operator. But In operator should return all those results even if...

thank you everyone, my problem has been solved. — tekf 23 mins ago
 
Saw that earlier. Did not click.
 
Ugh, I barfed from all the row pointers
 
If I were a moderator I would delete it in a way that no gods old or new would be able to see it
 
Hm, well it's a better question than I imagined
 
The data model renders any assistance as delaying the inevitable solution which is euthanasia
 
1:41 PM
Can't fault the effort they put in though
Must've read Erik's guide
 
16 hours ago, by J.D.
In the vein of silly and redundant code, I caught EF Core generating the following code after an INSERT today: SELECT MyIdentityColumn FROM MyTable WHERE @@ROWCOUNT = 1 AND MyIdentityColumn = SCOPE_IDENTITY(); instead of just SELECT SCOPE_IDENTITY(); lol
EF has a bunch of machinery related to making sure write queries only affect the expected number of rows, and that the right server-generated ID comes back.
There's a long discussion here about disabling it in scenarios where it doesn't really make sense: github.com/dotnet/efcore/issues/10443
 
2:01 PM
Peter seems to be having a tough time
 
> Another good reason to opt out of EF's optimistic concurrency is if you're using repeatable read transactions (or serializable/snapshot).
Repeatable read transactions are somewhat similar to optimistic concurrency, but managed by the database instead of by EF: the unit-of-work only succeeds if there was no conflict, otherwise an error is raised and the application has to retry.
> Serializable/snapshot transactions have overhead costs which vary across databases (e.g. may introduce additional locking), and also provide more than just checking that rows haven't changed (e.g. they guarantee a stable view of the database throughout the transaction).
I knew I'd regret following that link
 
Time to Make Some Friends©️
 
@ErikDarling this is the most unbelievable part
> I've already checked to make sure it's not being blocked, and that it's not the victim of some long running maintenance task, report, or other out of band process.
 
@Zikato hope burns eternal
 
2:19 PM
@ErikDarling Link?
It's an interesting point about instead of triggers breaking SCOPE_IDENTITY() though
That said, all bets are pretty much off as soon as you allow replacement actions
 
That deleted answer is reminiscent of other recent attempts
 
@PaulWhite I always wanted to go to cairns
Is Erik doing more friend making tweets?
 
2:37 PM
@PaulWhite Apparently, just one of those things: Apple: We didn't change anything. Is it working now?
 
@JohnK.N. Right.
@ErikDarling I see, thanks
This is the identity thing: dbfiddle.uk/QluvpeTV
 
Did you consult the DSM5 for identity disorders?
Also lots of questions around PSP I've been seeing
 
@SeanGallardy everyone is on pterodactyl or whatever now anyway
 
@ErikDarling Because Apple is removing Twitter from its app store?
 
i have no idea
 
2:45 PM
I don't twitter, I don't know if you're being serious or not
 
tiktok will stay until we all join emperor xi in the great american struggle session
 
I did see a bunch of headlines on the news this morning while at the gym about the apple twitter thing, that's about all I know
 
Some days the transcript makes me dumber
 
@PaulWhite 😂😂😂
 
2:47 PM
who says naming things is hard
 
> Rhamphorhynchus is a genus of long-tailed pterosaurs in the Jurassic period. Less specialized than contemporary, short-tailed pterodactyloid pterosaurs such as Pterodactylus, it had a long tail, stiffened with ligaments, which ended in a characteristic soft-tissue tail vane.
@PaulWhite Have you met Microsoft product names 😣
Their naming feels like a struggle session
inevitably it ends with being renamed 2-3 years later, with multiple iterations of such
 
When Gaia met Fido
Rhamphorhynchus is a great word tho
 
It sounded made up, I had to look it up for sure
 
@SeanGallardy Really? PSP in general or the complete solution offered in 2022?
 
Almost sounded like rhino copulation
 
2:50 PM
4 u
 
@PaulWhite Yeah, the implementation of said solution. There was a bunch of internal threads today about "how many plans" "what's the limit" etc.
 
🙄
 
I'm wondering how that will affect Erik's SPs
seems like query stuff so I tend to not care/stay away
 
I should check if random query variant ids made it into RTM. I don't think they did
 
I don't want that bad juju on me
 
2:51 PM
MULTI PLAN
 
ALL YOUR PLANS ARE BELONG TO US
 
@SeanGallardy He blogged about it being a disaster at one point idk if anything changed
It's very hard to care about everything
 
:( hmmm
 
I might enable trace flag 15210 just for the hell of it
 
Someone had an extra glass of wine today
 
2:55 PM
Pretty soon they're going to start spelling words
Mind you, 8008 was cool
It'll be a while before 80085 becomes available
 
@PaulWhite did i? i thought that was the other brent.
 
@ErikDarling Blogged in the transcript perhaps
 
hrm
well i don't like the query being detached from the proc in the cache
 
No, I'm p sure you did have a whinge
Yeah that was it
 
that's me
'avin a whinger
 
2:59 PM
quite right too
 
if i wanted that to happen i'd keep writing weird dynamic sql to fix parameter sniffing
ah well
can't "whinge" them all 🥁🥸
 
so they didn't improve it for RTM?
despite moderately influential feedback
 
honestly haven't looked
 
yeah, I'm also too lazy
 
been busy with deadlock and blocked process xml
 
3:01 PM
@JoshDarnell Fair enough, I realized that after ypercube replied. But again, couldn't the OUTPUT clause be sufficient here? And if they need the row count, then just SELECT @@ROWCOUNT in a subsequent query too. Seems silly to lock the table a second time after the INSERT to read back the identity value it already had access to, especially with an oddly crafted predicate with two functions in it. (I'm sure it's not normally a performance issue, but the potential is there.)
 
19 mins ago, by Paul White
This is the identity thing: https://dbfiddle.uk/QluvpeTV
@J.D. regarding output etc. ^
I really don't like the way it returns zero there
But what to do
 
I was about to qualify my previous comment with "as long as I'm not being an idiot...". But if I understand your example correctly, OUTPUT doesn't respect returning the correct identity?
 
It can't. There's literally no restriction on what an instead of trigger might do
 
Ah sorry, glossed over the trigger part lol.
 
I only included it because it came up on the github issue thread
Have to admire the ambition of the EF team
 
3:06 PM
 
did the cat also do that?
 
Only trigger worth mentioning
 
dba greg might need his license revoked
 
nemo could use some exercise
 
@ErikDarling Write bad triggers? I wouldn't put it past him.
 
3:07 PM
oh i thought that was a dead fish
 
looks p dead
 
@J.D. From my understanding, they can't just do SELECT @@ROWCOUNT because they are trying to confirm both things are true (only one row was inserted, and that row has the expected ID).
 
what percentage responsibility does the fish share for having such delicious colors
 
chip shortage is real
@JoshDarnell at the risk of exhausting your patience, what's a concurrency token in the EF context?
 
@JoshDarnell Idk, but I concede to bring an idiot for now after Paul's trigger example lol. Will spend more brain cycles on it during off-hours.
 
3:10 PM
tbf there's no good answer to the instead of example
even conceptually
 
Word up
 
@PaulWhite For the SQL Server provider, it's a timestamp / rowversion column.
 
@JoshDarnell Oh! That was my guess! I could've looked clever if I'd typed it out
Damnit
 
does the same problem exist with sequences as identities
asking for an idiot
2
 
@PaulWhite I fully expect you to rewrite history now 😂
 
3:13 PM
@ErikDarling Well the problem is, what to do if the original (say) one row insert was replaced in the INSTEAD OF trigger by a multi-row insert, or a delete, or a drop database...
 
That's the default and most convenient option, anyway. You can configure any column to function as your concurrency check (a datetime or uniqueidentifier for instance), but you have to make sure it gets updated whenever you make changes.
 
sure
 
@PaulWhite ... or drop a hammer in a head
 
Found some clustered indexes on GUIDs. Anybody?
 
@ypercubeᵀᴹ yep, instead of actions are just weirdness really
 
3:15 PM
Also shame on me for not finishing reading the transcript before re-commenting lol
 
comment addiction can be tricky
 
They don't limit you on StackOverflow, it's easy to get hooked...
 
Nov 22 at 2:42, by Paul White
Stack Overflow was a mistake
 
@PaulWhite shutdown with nowait;
 
a SOAP call to a microservice that nukes the query writer from orbit
 
3:17 PM
"INSTEAD of writing that insert trigger, quit your job and never touch a database again"
 
anything's possible really
 
I actually find it neat, how easy it is to do naughty things in SQL Server.
 
i always wanted something that would make a boxing glove on a spring pop out of someone's computer
 
can't beat the classics
POW!
 
or any three stooges gag
eye poke, seltzer, pie, nose grab
 
3:22 PM
@ErikDarling I imagined something that would send a shock through the keyboard - easier to implement and also blame on poor wiring
 
though pie and seltzer might encourage some
 
For example, the ability to literally order pizza from Domino's by executing SELECT Pizza FROM Dominos. It would be rather simple to create a CLR procedure that hit's Domino's ordering API. Then wrap that CLR procedure in OPENQUERY() and stuff it in a View called Dominos, like their stuffed-crust pizza.
 
May 3, 2012 at 21:23, by ConcernedOfTunbridgeWells
Good news. We shall have to teach her to say
"Daddy, when will they make an app so I can punch people over the internet?"
 
@J.D. That might be the most offensive message ever posted in this room
It even abuses an apostrophe
People do ask about SELECT triggers remarkably often though
 
3:27 PM
@PaulWhite 😁. I can do worse...
 
Please don't think of it as a challenge
 
Everything is a challenge if you're drunk enough...
 
i recently got drunk enough to forget i got lasik 9 months ago and tried to take out contact lenses
 
yikes
 
Lasik is the best
 
3:30 PM
👀 watch out for those retinas
 
except glasses make one look cool
 
you mean, keep an eye out for them
 
Erik was trying to recreate the 3 stooges classic eye poke...
 
i suppose there were some other crucial details
like i was very jet lagged and woke up somewhere in the night with very dry eyes
but also still rather tipsy
 
@J.D. Why not make it a CLR table-valued function?
 
3:38 PM
Needs more web services running the cloud
I watched a database attempt to undo 4 TB of log last night...
AFAIK it's still undoing
 
doesn't sound accelerated
 
That cloud storage pumping a whopping 20 MB/sec throughput
 
@SeanGallardy This little maneuver is gonna cost us 51 years
 
@Zikato Yeah it was really terrible to watch
I felt bad
That doesn't happen often
 
That's it, all this bad code talk is triggering me. I'm giving up and becoming a Mongo developer.
 
3:42 PM
Or.. and hear me out... you become a Mongo developer and believe you're the best code
 
It could have been worse. Erik could have multiplied it by 1000, and it would be 4 petabytes
 
oh, haha I read that wrong
I read it as "I'm giving up on becoming" not "and becoming"
Does 4 PB fit in cloud scale?
 
hyper scale
hyperscalescale
 
plaid scale
 
beat me to plaid
 
3:44 PM
great minds
 
ludicrous scale
Wonder if the rapper will have a say in that
 
I think he'll be coolio about it
oh shoot, apparently Coolio has died
 
Should have been wearing a TI vest
Got his at the 50 cent store
 
keep in mind sean has never listened to rap in his life
 
I know someone who has, though
does that count?
 
3:46 PM
the republican defense
2
 
better than the alternative defense, "someone somewhere may or in the future could hear this"
 
to be fair, I also don't listen to rap
 
Maybe I better czech myself before I wreck myself
I lakh rap experience
 
@Zikato because the goal is to do something poorly. 😉
 
@J.D. But why try that hard when Mongo is like RIGHT THERE waiting for you
 
3:50 PM
I'm surprised I haven't seen a mainstream SQL Server web services pack/tools
 
@Zikato I blame the hot ones
 
@J.D. Then add a Service Broker that uses a Linked Server to generate a dynamic SQL with synonyms
 
IIRC MySpace heavily used service broker
I don't think it's been touched since
 
@Zikato hey now, according to Paul 15 minutes ago, it's not a challenge...
 
I was just mentioning an example I've seen previously
 
3:57 PM
Lol goodbye... Hey bbaird, where's this train to Mongo and what time does it leave?
 
Anyone try any of the ultra wide curved monitors that can show up as 2 or 3 individual monitors in windows?
I'm thinking of consolidating my 3 monitor setup to a single one
 
i had a dell 38" uw that i moved off of when i got my crapbook
i don't think i'd go back, only because i really like having side monitors in portrait
 
I used a samsung 34" UW at work for a while. Really liked it
 
@J.D. same time as the crazy train
@SeanGallardy Not curved, but definitely using higher res larger screen for a while and prefer it to multiple monitors. If I had some extra cash I'd consider doing a 1440P ultrawide
 
4:14 PM
@bbaird oh good, so at least the music will be good.
 
4:27 PM
@ErikDarling Interesting, so mostly just due to the portrait mode? Any other issues or things you didn't like? I'm worried about screen sharing on a single monitor.
@bbaird Any drawbacks you've found?
 
@SeanGallardy That's been a problem for me on the receiving end. One of my coworkers has an ultra wide curved monitor, and very hard to see stuff when he screenshares. We work around it by either sharing just the app window he wants to share and resizing, or switching to a different monitor (he has a smaller one connected).
 
@SeanGallardy just the sharing thing @JoshDarnell mentioned
 
I remember somehow going around that using the fancy zones of PowerToys
 
4:48 PM
@SeanGallardy Try this one: youtu.be/2h4XozvqhJs
You can flip it from horizontal to vertical
Neck muscle training included in the price.
1000R (I think)
 
> This code was written by a tool
yes, yes it was
 
5:14 PM
@SeanGallardy yeah the separation is nice. most of the time i'm on a webex on my main monitor, and use side monitors for stuff that i want to keep an eye on without too much of a context switch.
with the uw that strategy was far less effective
 
I remember thinking PowerToys would be cool, but never could get into the habit of using the features.
 
my middle monitor is "pixel dense" or whatever and good for whatever other stuff
 
5:29 PM
Are you sure you want to post code with someone's actual email address in it on a public forum? — Erik Darling 1 min ago
some people eh
 
5:45 PM
@JoshDarnell That's my fear with the screen sharing stuff. I was wondering if it shows up as, say 3 hardware monitors, that you could then just "share" the one screen which would be normalish sized.
@ErikDarling This makes sense, so it was hard to cut up the UW to make it work like multiple different screens
 
@JohnK.N. oh my. That's.... wow.
That's a tv
that's uge
 
6:00 PM
only turtles could use that
 
@JoshDarnell has some nifty things....and though you don't use them all the time, when you need them they work great
 
 
1 hour later…
Another case of NoSQL "discovering" what maintenance scripts have been doing for years?
 
is there a maintenance script that cleans up indexes automatically?
 
I think it's pretty neat that it's built right into the product with a really excellent UI.
 
@JoshDarnell I'd forgotten about PowerToys. Thanks.
 
7:39 PM
@ErikDarling Would it be hard to do?
I know we like to be control freaks and "nothing changes without my say so" but if you did want this it's not a huge deal.
UI I'm indifferent
 
8:18 PM
@ErikDarling yeah, autovacuum ;)
@JoshDarnell the description looks like SQL Server's Missing Index suggestions but more clever (well, they claim it to be clever)
 
9:06 PM
@bbaird to merge overlapping indexes? not particularly
 
9:53 PM
@ErikDarling detecting them would be easy enough and usually only one index would be used generally (except in weird cases where one is covering for some reason while the other is not)
 
right, but merging key columns can be tricky sometimes, merging includes without overlapping key columns, minding filters and uniqueness etc.
idk, it's doable but programmatically annoying
 
10:22 PM
I doubt the RavenDB stuff is that complex - probably just looks to see if they start with the same few columns, takes the one used most often and jams all the columns in
 
I'd say the part that would be hard to get right is when to decide a merge or a removal of an index. Even if it checks how often indexes are actually used by queries, what if an index is only used by a monthly report and nowhere else .
@bbaird guess they didn't like calling it a bad idea ;) (Talking about the UDF on CHECK constraint question, not about RavenDB)
 
10:39 PM
@ypercubeᵀᴹ you might add my series to your list as well (if you want) straightforwardsql.com/posts/scary-scalar-functions-parallelism
 
11:31 PM
@bbaird right, I’d wanna be a little smarter about things, heh
 
11:44 PM
Like Erik could be any smarter
 

« first day (4348 days earlier)      last day (522 days later) »