😅 what year is this guy living in? SQL has scaled, probably better than any other software paradigm.
Yes, but if I do 80 joins and don’t specify join conditions, things get, like, super slow, Scoobs!
That’s why you should only left outer inner right join with a recursive select!
This guy is vibe coding his queries. That’s the future of programming :(
God I hate that term. They just put vibe in front of it and act like copy/pasting things you don’t understand is “coding.”
I think it is a perfect term. It so accurately describes how terrible the outcome must be. You can use it with other high-skilled jobs as well: vibe surgery, vibe piloting, vibe architecture.
vibe rator?
I love to hate this
Oh, he’s on AWS, that might explain some things
Y’all must be running Oracle DB if your SQL is not scaling.
Somebody needs to learn to design his database layout by common queries and performance standards.
Do we need to bring out the Mongo DB web scale video again?
First thing I thought of when I read this, and if you click through to the actual Linkedin post, that’s exactly what the twit is saying.
What is it about?
thanks for sharing this goldmine
How about adding indexes and stats on the appropriate columns, and have look at your execution plan while you’re complaining. Bushels of low hanging fruit are ready to be picked for devs who bother with a little SQL performance tuning.
And if doing pagination, order by an index and add a minimum bound, instead of doing LIMIT 5000,5000.
Oh and if the query optimizer picks a bad index for what you’re doing, you can force a better one.
And don’t do dependent subqueries and…
Idk, this is stuff I learned in my first year or 2 at my first job in the field. And I’ve had people on reddit tell me I’m a programmer not a software engineer because I don’t have the education so therefore I can only be a code monkey. Yet somehow Amazon pays people 5x my salary to not know how to use SQL efficiently while I’m out here building queries that show no non-index reads, sometimes in databases where I’m not even allowed to add new indexes because of legacy reasons.
But that’s work! How dare you ask me to do work!
So there was this QA tester filling a million documents in the database (NOSQL) and then using a single fetch to get all the data without pagination and then telling me my library was first being slow and then crashing.
Of course I didn’t handle
std::bad_allocas someone who has never coded in SQL and only knows some python, SQL is so scalable it sheds every 3 weeks
