Challenge #187: Generate Prime Numbers

A solution to last week's challenge can be found here!

This week's challenge comes to us from @Shreelatha! Create a workflow to generate all the prime numbers between 0 and 50. For this challenge, the number 1 is not considered to be prime as it only has one positive divisor.

challenge_187_start_file.yxmd

challenge_187_solution.yxmd

Preparation

Advanced

Data Analysis

Intermediate

Accepted answers

All comments

hlmess

Spoiler

Here's My solution. It involves comparing all combinations of numbers between 1 and 50 so it's probably not ideal if you wanted to identify primes up to a million or something but it works.

challenge_187_solution.yxmd

Jean-Balteryx

Here is my solution !

Spoiler

challenge_187_workflow.yxmd

David-Carnes

My solution worked great for numbers 3+. And when I say "great" I mean for small numbers, as this would fall into the trial division as referenced Prime number article in Wikipediam found in the text in the start file.

Spoiler

David-Carnes_187.yxmd

JamesFo

This one was fun, I've never thought of doing this before.

I wonder how many rows it could take before it started to struggle...

Spoiler

I suspect people will have a much cleverer way of doing this than me. Maybe i'll also try to work out a macro approach...

challenge_187_JF.yxmd

hellotf12

Here's my solution! Thanks for a fun challenge!

Spoiler

challenge_187_start_file_solution.yxmd

RolandSchubert

I used Fermat's primality test to identify prime numbers (test isn't absolutely reliable for very large numbers, but works fine for numbers < 341). I used 2 as a basis, so I had to check "2" separately.

Spoiler

challenge_187_solution_rsc.yxmd

OllieClarke

Here's my solution:

Spoiler

challenge_187_OC_solution.yxmd

cgoodman

Here's my solution. works quickly for primes 1 - 50.

Spoiler

Challenge_187_solution.yxmd

estherb47

A nice brain break for a Monday.

Spoiler

Challenge_187_EHB_completed.yxmd

JohnJPS

Spoiler

I used an old prime factorization macro I wrote a while back.
It finds all prime factors so is a bit overkill; but was trivial to reuse here.
Macros are pretty quick: this does 1 - 50 instantly and 1 - 10,000 in just over a second...
But... results diminish: it takes ten minutes to do 1 - 1,000,000.

c187_jps.yxzp

T_Willins

I reused most of the macros from Challenge #79, because recycling is good.

Spoiler

Iterative Macro Macro

Macro

Workflow

challenge_187_solution.yxzp

TonyA

Here's my solution. Not very different from the others.

Spoiler

I did optimize it a bit by looking only at numbers that were multiples of two and only using divisors less than half of the number. This cut out the first few primes (2,3,5) so I added those manually. But this is still an n squared algorithm so with every 10x increase in the number of candidates, the time increases by 100x. It's actually worse than that because for larger data sets we have to start swapping to disk which kills the execution time. This runs in less than 7 seconds for numbers up to 10,000. I did find a paper that discussed an algorithm of order n. I may try to implement that in Alteryx when I have some time. Maybe in 2025??

challenge_187_TonyA.yxmd

JORGE4900

I just upgrated to version 2019.3; I cannot wait to start using the Python interface to try this challenge. I remember doing this exercise for my Oracle database programming class.

Spoiler

challenge_187_Jorge_Solution.yxmd

TylerNa

Optimus prime would be so proud...

Prime Numbers.jpg

Spoiler

WC 187 Prime Numbers TN.yxmd

paulfound

Old favourite working out prime numbers. Up the number to 30,000 and got it ran in under a minute so that's good, won't break any records though. 🙂

Spoiler

Looks like I could improve around the append to summary section, this is bottle neck

Has anyone got any suggestions on increasing the speed? Wonder how the new Alteryx Engine 2 would get on with this? 🙂

challenge_187_solution_PF.yxmd

primes-to-100k.txt

patrick_digan

I did a little workflow on prime numbers for our last Indianapolis User Group meeting! I made use of @jdunkerley79 's formula add ins

Spoiler

First I generate numbers to check. I start at 2, then 3,5,7,9 etc. Then I generate the list of divisors (>2 and <=Square root) and check that it's not divisible. Then I use the filter tool to get the primes and iterate my counter. The cool thing is that the counter from the filter tool actually feeds back into the generate rows tools and is the condition that stops it. This is possible with the VarNum and VarText functions that James created. They're so powerful! You will need to change my file from yxmd to yxwz.

Annotation 2019-10-28 154345.jpg

Challenge 187_Digan.yxmd

hbraunius

My solution attached.

challenge_187_solution_hb.yxmd

justindavis

Spoiler

challenge_187_solution_justindavis.yxmd

Adam_Dooley

Only used 5 tools (aside from the Expect Equal macro) so that's neat. It helps that we only had to go to 50 or else this would get too manual.

Spoiler

challenge_187_AD.yxmd

meeravijayan2011

Spoiler

SPOILER

challenge_187_solution_file.yxmd

Ladarthure

My solution, fun challenge!

challenge_187_start_file.yxmd

RWvanLeeuwen

Here's my take on the matter (I considered 1 to be a prime).

Spoiler

challenge 187 - RWvL.yxmd

Martyn

I decided to take a more dynamic approach and used an iterative macro.

It works for checking numbers up to 1 million (it does 1000 iterations for this as that is the square root of 1 million).

Spoiler

challenge_187_completed.yxzp

sprodger

Spoiler

challenge_187_solution_SP.yxmd

SamDesk

Spoiler

Sam

challenge_187_start_file.yxmd

kat

Fun! 🙂

Spoiler

Challenge #187.yxmd

dhtay

Here's a relatively straightforward solution.

However, it's not very feasible if we were to start with a larger number (>10000).

For scaling up, a better solution might be to use a batch macro that tests the divisibility of each new number by previously found prime numbers. I'm not sure if that's possible though.

Spoiler

challenge_187TDH.yxmd

RoDO

My solution

Spoiler

challenge_187_RODO_Solution.yxmd

unknown

challenge_187_start_file.yxmd

Flo_G

Very interesting challenge - I'm sure everyone went about it in a different way. Here's mine:

Spoiler

I calculated the root of the input number, 50, to get the list of integers to use for testing - 2 to 7. I created the list of numbers to be tested, 2 to 50. I excluded the single-integer prime numbers for simplicity. For the others, I checked which ones have a remainder when divided by one of the numbers between 2 and 7. I joined the two lists for a full list of integer numbers between 2 and 50.

One step includes removing numbers equals to the dividends, this makes the workflow usable for very large numbers as well. It's still quite fast with numbers in the thousands 🙂

challenge_187_fg.yxmd

lmorrell

A fun challenge for sure! Great to approach these fundamental concepts through the lense of Alteryx.

Spoiler

challenge_187_LM.yxmd

Kucsi

Hi Christine,

please find attached my solution.

Spoiler

Thanks,

Kucsi

challenge_187_Kucsi_solution.yxmd

PaulRB

@paulfound Sounds a little slow. My solution generates primes up to 1,000,000 in 55 seconds.

(I did make a small optimisation to get it under 1 minute which means it does not find 2 as prime, but does find the 78,497 others).

JohnJPS

Did it again, spot-checking the first several primes in Alteryx, and using an R package that can handle very large numbers... I ran it again on 1 - 1,000,000, and also added in a bunch of very large numbers for kicks.

Spoiler

These are the added numbers:

1234567890
12345678901
123456789012
1234567890123
12345678901234
123456789012345
1234567890123456
12345678901234567
123456789012345678
1234567890123456789
12345678901234567890
123456789012345678901
1234567890123456789012
12345678901234567890123
123456789012345678901234567
1234567890123456789012345678901
123456789012345678901234567890123
1234567890123456789012345678901234567
12345678901234567890123456789012345678901
170141183460469231731687303715884105727
20988936657440586486151264256610222593863921

With the spot checking, it's down to about 25 seconds or so.

c187R_jps.yxmd

matthewbechard

Here's my 1st published solution!

I generated 50 rows, appended it to itself. I then used a formula to divide both sets of rows, split the result so I could filter on rows that have a .00000 in the decimal, and did a few other tricks.

Weekly Challenege #187.yxmd

SeanAdams

Very elegant @RolandSchubert !

cmoussa

challenge_187 CTM.yxmd

Quick Links

This months top contributors

atcodedog05 19458

Qiu 15867

binu_acs 15708

MarqueeCrew 13708

apathetichell 13703