Markus Winand's Recent LinkedIn Posts

Order-equivalent OVER clauses https://lnkd.in/dpEP-wrC About an important but rarely appreciated aspect when using multiple OVER clauses in a single query.

AI is the future? Well, if we cannot buy computers anymore, it sounds like stone age to me.

My next SQL Performance Training starts on 21st April, at 13:00 CEST (~ time of this post). This and more trainings on my website: https://lnkd.in/ewG8eBN3

What does this query return? SELECT 1 AS a FROM (VALUES (1), (2)) t(a) GROUP BY a One or two rows?

Order-equivalent OVER clauses https://lnkd.in/dpEP-wrC About an important but rarely appreciated aspect when using multiple OVER clauses in a single query.

“We don't have a scaling problem. We have an assumption problem.” Exactly. During the NoSQL hype I used to say that those systems solve problems everybody would like to have - but hardly anyone actually has.

SQL Comments, Please! https://lnkd.in/dKJxaqPf Standard SQL offers two flairs of comments: one starts by two dashes (--) and goes to the end of the line. The other starts with slash-asterisk (/*) and ends with asterisk-slash (*/). That might remind of comments in other programming languages. Unlike many other programming languages, the second form allows for nested comments in SQL. This is particularity useful when commenting out code that contains /*…*/ comments.

Order-equivalent OVER clauses https://lnkd.in/dpEP-wrC About an important but rarely appreciated aspect when using multiple OVER clauses in a single query.

”Complexity isn't a sign of maturity. Simplicity is.“

But first of all: triple check traditional tuning. Indexing, making better use of SQL features. Let’s not skip these options.

Impressive, yet practically useless. After reading the linked article, this is my resume about the new C compiler (almost) entirely built by AI within two weeks consuming token worth $20.000. This article looks behind the marketing and benchmarks the compilation as well as the compiled output against GCC (also SQLite ;). I'd say it is a must read. https://lnkd.in/dSh7iwC3 IMHO the question is this: is this one of the infamous 80% solutions where the remaining 20% turn out to be actually rather 2000%?

Ultimately, LLM based tools are still tools. So what do you think?

Well…very small sample.

"AI engineers" only consider the explicit requirements (if at all). Human engineers read between the lines and implicitly take care of implied requirements (ideally, actually not all do so).

Nowadays, convenience beats everything. Sex sells is so last Wednesday. Today, convenience sells.

Read Skew

Ultimately, LLM based tools are still tools. So what do you think?

I wonder when we see the first OpenClaw worm, if there isn't one already. Switching off your Mac Mini won't stop it.

While Martin focuses on management in recent years, he has a technical hands-on history from his DBA times and from translating my book "SQL Performance Explained" to Spanish :)