« first day (2726 days earlier)      last day (2146 days later) » 

2:45 AM
13 years of not a real database
 
 
2 hours later…
4:25 AM
-1
Q: Tag rename aggregate -> function-aggregate

Evan CarrollInspired by this, Separate tag for [set-returning-functions] We have another one that we may want to clean up, [aggregate]. This seems like it should be function-aggregate. As far as I know, nothing is an aggregate except for a function, it's also a lot harder to find like that.

 
5:08 AM
@EvanCarroll flag it
 
 
2 hours later…
6:49 AM
Why is this question closed? dba.stackexchange.com/a/209975/2639
This kind of shit gets old. That in every way is on topic with the /faq
 
7:43 AM
This utter nonsense has gone viral today: codeburst.io/…
 
@EvanCarroll because you shouldn't use tabs /me ducks
 
@EvanCarroll Enough people with the necessary powers deemed it off-topic
But you know that already
 
@dezso you're safe though bruh =)
 
> This article will certainly rank near (if not at) the top of my list of totally uninformed and misleading technology related writings I have ever read in the last 30 years.
Could not agree more. Except I haven't spent 30 years with reading tech articles yet.
 
That article is horrible.
Ahahha Edwin Buck at the top. He goes to Houston Linux Meetup Group
how do these cats get phds
Seems all of his articles are that bad.
 
7:56 AM
@EvanCarroll I've seen more than one PhDs without any recognizable cleverness. Some of the totally undereducated neighbours on the wine hill (that's Hunglish, beware!) were much brighter than these.
 
Ben Garvey's blog is awesomely bad.
if you want more amusement.
 
@EvanCarroll the OP states in the answer to it that he was already producing enterprise systems in the 70's. Based on this, he should go back learning or just retire.
@EvanCarroll whoz dat?
 
another idiot who is trying to make it in the blogosphere.
 
8:24 AM
@Philᵀᴹ Oh ,we criticized him a few days ago for not offering an alternative, but he has since posted a followup i see
He's rebuilding Stack Overflow in a much more sensible way :D
 
@EvanCarroll Half of my management got into Oxbridge. Science is powerless to explain how they did it.
I have never been so amazed at how porous the admissions process at elite universities really is.
 
The real disservice is that he sells stupidity to stupid people.
 
@EvanCarroll Just like my management team then.
One shark and a confederacy of dunces.
 
His audience can't be actual DBAs. He's talking to people who are coming into the industry and trying to learn and telling them that they don't have to because it's not the 1970s.
speaking of dunces, can I get some reopen votes ;) dba.stackexchange.com/questions/209974/…
I already solved the issue.
 
8:42 AM
@Philᵀᴹ What a load of horse manure.
@TomV gotta be a 10x-er
 
@TomV that's a... a...
 
9:28 AM
@TomV looking forward to seeing that stackexchange_derivate running
 
 
3 hours later…
12:35 PM
@MichaelGreen ^^
 
 
6 hours later…
7:39 PM
so, some third-party install has created a database with 18MB of data, in simple recovery model, and has blown up the log to 18GB so far. I'm thinking something is definitely not right. The transaction causing the log to grow just keeps doing the same thing over and over again. Must be an infinite loop. Gotta love vendor software.
 
8:09 PM
Does this smell like bad code to anyone else? This is in a where clause from the same vendor:
WHERE FLOOR(ObjectID / 2) - (FLOOR(FLOOR(ObjectID / 2) / 2) * 2) = 1
unless I'm missing something, it looks like that will always be false to me.
ObjectID is an int.
I guess it can be 1, so maybe it's just some magic number kind of thing.
 
Since ObjectID is an int, we can ignore the FLOORs. Then we can replace Object / 2 with an X. So now we get X - X / 2 * 2 = 1. Can this be true? Yes, it can. It'll be true when X is odd
 
not for 1
so weird. ObjectID is the primary key for the table.
 
Not when ObjectID is 1, but when ObjectID / 2 is 1
 
Maybe they have two threads, one for the even ObjectID rows, and another for the odd ObjectID rows.
seems you could easily do WHERE ObjectID % 2 = 1
 
It's not about even and odd ObjectIDs but about even and odd pairs of consecutive ObjectIDs
@MaxVernon ObjectID / 2 % 2 = 1, more specifically
 
8:19 PM
oh well, I guess without seeing the code, I'll never understand why they are doing that.
plus it's Friday afternoon.
 
For 0 and 1 the left side will yield 0, making it false, then 2 and 3 will give you 1, thus true, and so on
 
9:05 PM
@MaxVernon what @AndriyM said. In other words: Take ObjectID. Divide by 4 (integer division). If the remainder is 2 or 3, the result will be TRUE. If it is 0 or 1 it will be FALSE.
 
754 upvotes and I think this answer wrong lol dba.stackexchange.com/q/210408/2639
That's insane.
That comment correcting it has 108 upvotes lulz.
 
I've never seen an advertisement to downvote one's own post ;)
 

« first day (2726 days earlier)      last day (2146 days later) »