Kata

The code gives out the final list with the Harshad numbers but it gives the execution runtime error. Please help.

👉😎👉

I'm afraid there is not much to do about it... :-(

Have you tried pouring coffee into your keyboard?

With the information you give, that's about the best advice I can give. That should make them perform the same again.

Seriously: post your code ( marked up and spoilered ). Right now, we don't even know what language you're using.

Have you tried pouring coffee into your keyboard?

LOL.

Should be fixed. Don't forget to reload.

Is it fixed?

Yeah, I hope so too! :ROFL:

Working on JavaScript now.

Yeah, I hope so too!

I guess not

I developed with Integers ( and bigger numbers ), then changed all Integers to Ints - but I forgot the initial code.

And I only tested the initial code against the example tests.

Yes, the limit could be set higher (at least 10^20, easily), but the idea is to exclude naive solutions. It may be a bit easy for 5 kyu, although there is not much difference between 5 & 6, right?

Thanks for the compliment, and keep thinking! :-)

Then you need to change your approach, as it's (easily) doable for even bigger numbers.

I find 5 kyu about right. Oh, and thanks for the compliment!

I think this is a tough kyu 5 ;)

As I explained below already, the result of the truncation has to be a Harshad number (not a right-truncatable Harshad number). If you can "successfully" repeat the truncation (=the result is a Harshad number), then it's a right-truncatable Harshad number.

For me it sounds clear (I took the description directly from Project Euler), but help me reformulate it to make it more clear.

It doesn't really matter if the result doesn't have to be truncatable, all truncations of 1-digit numbers are Hashad numbers.

This also means that 1-digit numbers are not (right) truncatable, because the result is not a number.

What exactly is not a number? It can't be truncated at all, so there's no result of a signle truncation at all, thus the result of recursive truncations is an empty sequence and it contains only Harshad number.

OK everyone, how about this?

A Harshad number (or Niven number) is a number that is divisible by the sum of its digits. A Harshad number that yields only Harshad numbers while truncating (removing) the last digit, is called a right truncatable Harshad number.

For example 201 is a Harshad number. Truncating the last digit gives 20 then 2 which are also a Harshad numbers. (2 cannot be truncated anymore.) Thus 201 is a right truncatable Harshad number.

Is it different somehow from what it was? Replacing 201 with 2: ... 2 is a Harshad number ... Truncating... (2 cannot be truncated anymore.) Thus 2 is a right truncatable Harshad number.

I believe yes:

A Harshad number that yields only Harshad numbers while truncating (removing) the last digit...

Can you truncate 2? Does it yield a Harshad number?

You know, you can just limit the lower bound to have 10 <= a, and this wouldn't be relevant anymore ;-)

Can you truncate 2? Does it yield a Harshad number?

One time - questionable (0 is a weird number that acts like a 0-digit number, but is often represented by 1 digit), recursively - you can and it results in a sequence of Harshad numbers.

How is it resolved??? If it's literally "always results in a Harshad number", then no number is truncatable because there's 0 in the end. Assuming that it applies to the numbers before 0 exclusively, 1-digit numbers are truncatable because there's no non-Harshad numbers in the sequence.

mmmm... correct.

[...] a Harshad number that, while recursively truncating the last digit, always results in a Harshad number [is] a right truncatable Harshad number

Well, if you truncate a 1-digit number, there is nothing left (~ an empty "string"; not zero!). As "nothing" is not a number, you don't have to check/worry, if it's a Harshad number or not...

Description updated.

As "nothing" is not a number, you don't have to check/worry, if it's a Harshad number or not...

Correct, but this exactly means that a 1-digit number is truncatable. All numbers in the sequence are Harshad numbers (because the sequence is empty), in other words there's no non-Harshad number in the sequence.

You either have to define single-digit Harshad numbers as right-truncatable or you have to redefine right-truncatable Harshad numbers. As they are defined now, there are no right-truncatable Harshad numbers, because any number, right-truncated enough times, ends up being a single-digit number; which you have defined as non-right-truncatable.

Consider 201, given as an example in the description:

201 % 3 == 0 #Harshad number
20 % 2 == 0 #Also Harshad number
2 #Also Harshad number, but not right-truncatable (RT) by definition;
#Therefore 20 is not RT, because it's recursion is not RT
#Therefore 201 is not RT, because it's recursion is not RT

The result of the truncation has to be a Harshad number (not a right-truncatable Harshad number). If you can "successfully" repeat the truncation (=the result is a Harshad number), then it's a right-truncatable Harshad number.

For me it sounds clear (I took the description directly from Project Euler), but help me reformulate it to make it more clear.

Let's call a Harshad number that, while recursively truncating the last digit, always results in a Harshad number a right truncatable Harshad number.

This is the part that lead to confusion for me. I can see why it means what you say, but it wasn't my interpretation. On Project Euler, the last bit is italicized, which I think helps set it apart as a definition. At the least, a comma should be added, but a rephrasing may be in order.

We are really talking about recursive Harshad numbers ending with a single-digit number (which are Harshad numbers, but we can't just stop recursion at any Harshad number). Arguably the definition allows us to recurse "past" the single-digit number into undefined territory. The end point might as well be explicit.

So you might say something like:

A right-truncatable Harshad number is any Harshad number which, when recursively right-truncated, results in a Harshad number at each truncation until it ends with a single-digit Harshad number. Single-digit Harshad numbers are defined as non-right-truncatable.

I think this jibes with your definition and is also explicit about the end point, while avoiding ambiguity (or even caring) about how single-digit Harshad numbers are defined.

But, I will also add that defining single-digit numbers as non-truncatable is not particularly useful, makes the recursion weird (we look for Thing until we find Another Thing Entirely), and is not how other truncatable numbers are defined. Note that 2 is a truncatable prime[0] in both directions. So my true suggestion would be to define single-digit numbers (other than zero) as right-truncatable Harshad numbers.

[0] https://en.wikipedia.org/wiki/Truncatable_prime#Lists_of_truncatable_primes

Thanks for your feedback. Description updated, stating that 1-digit numbers are excluded.

I left the high values in the sample tests, because you have to deal with even higher numbers in the random tests;

Yes, but the main interest of the random tests in this type of kata is to get a wuick check of your code. So maybe comment out that test so that the sample tests are runnable without perf requirements, but the user still gets the info about the expected range?

Thanks for the approval!