I don't know.  Setting STATISTICS ON doesn't work on Fiddle

Simplified without the 4th CTE

declare @max_number       int = 100;
declare @half_max_number  int...

• This reply was modified 2 years, 5 months ago by Steve Collins.

As one query

DECLARE @maxNumber INT = 1000;

WITH BitArray AS (
    SELECT [value] AS Id, 1 AS Bit FROM generate_series(1, (@maxNumber - 1) / 2,...

Aha the recursive CTE has been replaced with a 3rd usage of the GENERATE_SERIES function

https://dbfiddle.uk/bhPaKD5f

-- Create a table to store the bit array
CREATE TABLE BitArray (
 ...

More from ChatGPT-4

Now, let me explain how the Sieve CTE works and how it fits in with the rest of the query.

The Sieve of Eratosthenes algorithm works by iteratively marking...

After giving it a few tries lately I've been trying to get GPT-4 to do things which wouldn't otherwise be attempted.  So here is AI's implementation of the Sieve of...

Ok the proof by induction is acceptable

If 2 doesn't add or subtract from what's being measured that's a good reason to get rid of it imo

For my projects the upgrade...

including p.N%6 IN(1, 2) It does increase the performance a bit, so I would argue it is appropriate. It changes the number of prime candidates from n/2 (odd numbers)...

Shifting gears, there's a slight enhancement to Steve's algorithm that reduces the execution time from 2 seconds to 1.8 seconds. This improvement involves testing whether the prime candidate is...

That's the article which had shaped my low expectations.  Good to know it's been improved.  It's also interesting Jonathan used GENERATE_SERIES in place of the non-itvf method of TOP and...

These are great comparisons.  The so-far winning code was jointly arrived at.  Interesting in a straight comparison GENERATE_SERIES appears faster than fnTally

At the moment I don't have access to an instance of SQL 2022 to test with.  That could change in the near future tho.  For now I was just fishing...

Why mess with fnTally when what we really need is for Jeffrey to rewrite the query?  Ha, then we also get a baseline of the functions.  Also how does the...

Each number you test for primality needs O(sqrt(n)*n) calculations, so it will slow down for larger numbers. Also, if you increase @n to 10 times its previous value each...

You only need to test up to the square root of a number to test for primality, which is quite a saving:

declare @n   ...

Oh you're right.  This is not the final version but I'm not finding my notes on this.   Iirc if you follow the links it leads to a SQLlite script with...