for else clear solution for Three Words by veky - python coding challenges

Enable Javascript in your browser and then refresh this page, for a much enhanced experience.

for else solution in Clear category for Three Words by veky

def checkio(words):
    succ = 0
    for word in words.split():
        succ = (succ + 1)*word.isalpha()
        if succ == 3: return True
    else: return False

Feb. 20, 2014

Comments:

Cjkjvfnby on Feb. 26, 2014, 7:40 p.m.
You treat boolean as int, it does not deserves it. :)

veky on Feb. 26, 2014, 8:02 p.m.
Liskov says it does. :-P <pre class='brush: python'>>>> issubclass(bool, int) True </pre>

maurice.makaay on Nov. 3, 2014, 7:33 p.m.
Liskov wouldn't see the difference ;)

veky on Nov. 4, 2014, 5:37 a.m.
Exactly. So why do you think you see it? :-)

akeelmido on Aug. 22, 2021, 9:41 p.m.
I've enjoyed your solutions but would have to agree, that despite bool being a subtype of int (which makes sense in the context of python), it would not pass code review using it in a numerical context (for newbies that are interested in the topic)

veky on Aug. 22, 2021, 9:51 p.m.
This is not about Python. People have treated <a href="https://en.wikipedia.org/wiki/Iverson_bracket">truth values as integers</a> (especially multiplying by them) before Guido even started going to school. :-)

akeelmido on Aug. 23, 2021, 12:08 a.m.
Sure, I started with C and it's the same there, but in this case the context is python and its history is my only point. The ultimate point being The intention isn't to replace 1+1 with True + True. Hence why the use of bool in a numerical context would fail a code review.

veky on Aug. 23, 2021, 12:12 a.m.
Then I don't see what your argument is. It seemed you were saying that this (bools being ints) is some kind of accident, and not a fundamental mathematical fact. It is in fact the latter, and it has nothing to do with Python itself. "Newbies interested in the topic" also include people who should know about it. This is not about Python history. It's as if you say that in Python, 2+3==5 because of history. I mean, yes, it is, but there is also a deeper reason. ;-)

akeelmido on Aug. 23, 2021, 12:17 a.m.
My point being that using bools in numerical contexts is not generally accepted as proper coding in the real world. It's a distinction that ultiamtely doesn't really need to be made in checkio.org solutions, but the thread went there so here I am.

veky on Aug. 23, 2021, 12:27 a.m.
I really don't see where do you draw the "real world" line. Do you consider Iverson's articles as belonging to "real world" research? Is Knuth "real world" enough? Or do you only trust the BDFL? Then please look at <a href="https://stackoverflow.com/a/6865824/1875565">this</a>.

akeelmido on Aug. 23, 2021, 12:36 a.m.
Real world as in companies and the code reviews they perform. It has nothing to do with Guido or any language making use of bools as ints. If you use a bool to perform arithmetic, your code won't be accepted and you will be asked to modify it. This line in particular would not pass muster: succ = (succ + 1)*word.isalpha() Not saying it's bad I'm just saying for people that are learning to code that out in the professional setting it doesn't fly

veky on Aug. 23, 2021, 12:40 a.m.
How do you know? Have you ever done it? I hate that I'm doing this, but I have to pull the rank now. How many years of experience have you working in the industry, programming Python professionally, and how many code reviews have you done? (Guido has been at Dropbox for 6 years.)

akeelmido on Aug. 23, 2021, 12:46 a.m.
Three years, but python only one, not that the language makes any difference in this discussion as it is a general coding issue. I actually haven't been involved in code reviews that failed due to those reasons, it was others who were using languages other than python. One was C and the other I think C# but I can't recall. The reason being that using a bool to perform arithmetic is not clear, slick, yes, but not clear. Code needs to be readily understood by human eyes. As Knuth put it: "humans, not computers, read programs"

veky on Aug. 23, 2021, 12:58 a.m.
Then I have more experience in doing Python code reviews professionally than you. And one of them was about a similar issue (including an item in a list if a bool was True). It was an interesting discussion, but in the end we concluded that one line is better than three, the step was elementary, and factoring it out into a function would just need to transfer too many arguments and result in a shallow abstraction. <blockquote> As Knuth put it: </blockquote> Yes, and many other people (Abelson was the first, IIRC), including Guido. And, look, all of those people are perfectly fine with the idea that truth values are integers. Read Concrete Mathematics some time. <hr /> And don't deceive yourself, of course language matters. Bools being naturally embeddable into integers is a mathematical fact independent of language, but languages differ regarding the degree up to which they incorporate mathematical facts. [As far as I know, C has still not made up its mind about pow(0,0).:]

akeelmido on Aug. 23, 2021, 1:07 a.m.
Language has no baring on this issue because it's an issue of human readability. Many languages are capable of many things that are prohibited in project settings. When you have a massive project, you don't want code that will be confusing. Including an item in a list if a bool is True is NOT an instance of using a bool arithmetically. The code you posted above is an example of that.

veky on Aug. 23, 2021, 1:12 a.m.
How is <pre class='brush: python'>self.payments = [some weird comprehension] + [rest]*include_end </pre> not using a bool arithmetically? <hr /> Readability criteria themselves are language-dependent. What's readable in LISP might not be readable in Haskell (if directly translated), and vice versa. I think a much better word is "idiomatic" (instead of "readable"), and then it's obvious that it does depend on the language. <hr /> Ok, I think this discussion has reached the end of its usefulness. But I think it has been a productive one, and I've certainly learned something.

akeelmido on Aug. 23, 2021, 1:17 a.m.
<blockquote> Ok, I think this discussion has reached the end of its usefulness. </blockquote> Agreed have a good one!

akeelmido on Aug. 23, 2021, 12:57 a.m.
I do still enjoy your solutions to the problems I've done so far!

odwl on March 5, 2014, 5:02 p.m.
I was sure you would have found something in l.4:) I do agree with the comment that treating boolean as int is less good than an equivalent without. I did the following by using bool as bool count = count + 1 if w[0].isalpha() else 0

veky on March 5, 2014, 5:21 p.m.
You're making yourself jump through hoops without any reason. I hate when people deliberately tie their hands for some superstitious reason. One hates exceptions because he's been told they are slow (in another language!), other thinks bool being subclass of int is somehow impure, the third writes every function as a class because "it's more general", fourth wants to banish every lambda... nonsense. Language is language, with all its features. At least Guido is not afraid to admit if something is a genuine design error (and there were quite a few, between Py2 and Py3). This is not. bool being subclass of int is amazingly useful, in so many contexts. Read Iverson sometime. ;-) <a href="http://en.wikipedia.org/wiki/Iverson_bracket">http://en.wikipedia.org/wiki/Iverson_bracket</a>

TomTerebus on Aug. 19, 2014, 6:17 p.m.
While your solution might not be how I would do it, your insight, resources, and explanations are spot on. Thank you for sharing!

veky on Aug. 20, 2014, 5:55 p.m.
Thank you very much (and for +es on the other comments;). Just for the record, my intention is not really to write a solution like someone else would do it. I'm here to show people various reasons why Python is amazing, and if it takes using some relatively unknown aspect of the language, I'm cool with it.

Marcin_Coles on Oct. 31, 2017, 7:58 a.m.
And that's why I always look for your solutions :)

Freiling on July 1, 2017, 11:44 p.m.
Yeah, I could read this dude all day.

veky on July 2, 2017, 7:39 a.m.
I'm in the Land of Weird Compliments again... :-D

Cjkjvfnby on March 5, 2014, 9:24 p.m.
IMHO using bool as int is ok. In my simple solution I use if/else block, because it is more readable. In oneliners I refer other style: <pre class='brush: python'># and I don't want untie 'or 0' succ = word.isalpha() and succ + 1 or 0 </pre>

veky on March 6, 2014, 2:43 a.m.
Now you're obviously inconsistent. :-D But that's no news. Just a little question: how would you feel if someone wrote <pre class='brush: python'>b = a + 1 if a != 4 else 5 </pre> , claiming that number 4 is somehow special, and should be handled in a separate branch? That's how I feel when I see above code. :-]

Cjkjvfnby on March 6, 2014, 6:54 a.m.
<blockquote> Now you're obviously inconsistent. :-D It take me a lot of time to find this path :) </blockquote> I will feel good. I am always feel good. :)

dfridman on April 26, 2014, 11:35 p.m.
I'm just learning and I usually come to the solutions to see how more experienced people do it. I often look at comments in these examples for further explanation/insight, but often times there is not much info. So, I just thought I'd annotate this with some of my key takeaways from this, to help out the next person. <ol> <li>Boolean can be treated as integers (true=1, false=0)</li> <li>'Return' acts as a break in for loops (and other such operations)</li> <li>You can use else statements with for loops. They are executed when the for loop ends naturally (as in not from a 'break')</li> </ol>

veky on April 27, 2014, 2:43 a.m.
Basically yes. <ol> <li>bools are ints. isinstance(False, int) returns True.</li> <li>Yes, return is kinda like raising an exception (it's not really implemented that way, but conceptually works similarly).</li> <li>Yes. Here it isn't really necessary, since just "return False" in line 6 would do the same thing - but I wanted to use it for education value. ;-)</li> </ol>

Peter.White on March 26, 2015, 9:10 p.m.
<blockquote> <ol> <li>You can use else statements with for loops. They are executed when the for loop ends naturally (as in not from a ‘break’)</li> </ol> </blockquote> Oh, missed that one. Read it like it was in an if statement. Thanks.

veky on March 27, 2015, 1:54 a.m.
The point is, the analogy with `if` is the easiest when paired with `while`: <pre class='brush: python'>if cond: while cond: do something do something else: else: do something else do something else </pre> They are both saying: cond being true means something will be done, cond being false means something else will be done. The only difference is that with `while`, after something is done, we are back at the beginning, and with `if`, we are at the end. Of course, `break` goes to the end unconditionally, `continue` goes to the beginning unconditionally. Now the analogy is extended to the `for` loop just by analogy of `for` with `while`. It's interesting because the `for` loop internally works by catching `StopIteration`, but the `try`...`except` mechanism also has the possibility of using `else`, in which case it means the opposite thing: the suite is executed if no exception has been `raise`d in a `try` suite. If you interpret all the `except`s as one big `if BaseException is raised`, then that `else` makes sense.

dalluca on Feb. 28, 2016, 5:40 p.m.
Woowww ... this is the right way to use comments!! Thanks :-)

veky on Feb. 29, 2016, 4:56 a.m.
That's what I'm telling you. <a href="http://www.checkio.org/mission/buildings-visibility/publications/veky/python-3/closure/share/c1ed9f3521299ca8cac37d5c1bdcda28/#comment-4368">All the time.</a> :-)

tilikum on June 12, 2014, 11:32 a.m.
I honestly do not understand how does this counts 3 words in a row, and not just 3 words.

veky on June 12, 2014, 11:47 a.m.
You see that "*word.isalpha()" over there? If word.isalpha(), then succ+=1. Otherwise, succ=0 anew. Ok?

Jay_Jay on Sept. 29, 2014, 12:58 a.m.
Wow that is smart... because if you multiply anything by true it stays the same, anything by false it becomes zero again. Another great one, I was initially confused but this made it click, thanks!

veky on Sept. 29, 2014, 6:45 a.m.
It always bugs me when I read such comments, it seems to me as if people think such things are "coincidentally" true. Same as when a 5 year old figures out that 3+4 and 4+3 are the same, they go "wow, what a coincidence!". :-) Read about Iverson bracket some time. There's a whole theory about embedding bool into int.

tilikum on June 12, 2014, 12:08 p.m.
Nope. word.isalpha() returns a boolean, how can you even multiply this? Second, if you have a for-loop, isn't it suppose to run for every element? aren't you missing an 'if' statement?

veky on June 12, 2014, 1:07 p.m.
You can multiply str and list and tuple, why not bool? :-] But really, this is even much simpler than those other multiplications. bool is a subclass of int, which means bools are ints (and multiply as them). Look: <pre class='brush: python'>>>> 5 * True 5 >>> 7 * False 0 </pre> if statement is not needed here (since the target, succ, is the same). Maybe you could argue that if expression would be appropriate here, but my opinion is that here it's also too complicated, with nonstandard evaluation and other hairy stuff. This just says: add 1 to succ, reset down to zero unless word.isalpha(), and name it (new) succ. Simple arithmetic. :-)

lesenv on Nov. 18, 2014, 10:55 p.m.
haha, NOW i get it. thanks

veky on Nov. 19, 2014, 4:12 a.m.
<blockquote> Now is better than never. </blockquote> <pre class='brush: python'> -- Zen of Python </pre> ;-)

lesenv on Nov. 30, 2014, 12:12 a.m.
how do you markup your code? like with the border around "-- Zen of Python"

veky on Nov. 30, 2014, 5:27 a.m.
See those icons in the toolbar right above the comment box when you write a new comment? Mouseover them, you'll learn a lot. ;-)

lesenv on March 12, 2015, 10:41 p.m.
thanks, i learned <pre class='brush: python'>a lot :) </pre>

veky on March 13, 2015, 5:53 a.m.
And so we complete the <a href="http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-20867">cycle</a>. :-)

tilikum on June 12, 2014, 1:45 p.m.
10x for the explanation!

tigelane on Jan. 2, 2015, 4:56 p.m.
and thanks for asking the question. :)

veky on Jan. 3, 2015, 9:02 a.m.
This is really important. We are all learning here. Newbies are learning Python, I am learning what interests them. I wouldn't be able to do so without your questions. So keep asking them. And if you want me to review some solution of yours, just drop me a link.

Aiadan on July 9, 2014, 1:15 p.m.
I find this to be a creative solution but not very clear.

veky on July 9, 2014, 7:21 p.m.
What's creative there? And more importantly, what's not clear? You have a finite state automaton that reads word by word and remembers how many .isalpha()s it has seen in a row. If it ever reaches 3, it returns True, else it returns False. If you just mean using bool arithmetically, please see other comments here.

veky on July 9, 2014, 8:35 p.m.
Just by giving minuses you won't accomplish much. Speak your mind.

Cjkjvfnby on July 10, 2014, 7:20 a.m.
I got email that you arguing with yourself ([CheckiO] veky has replied to veky's comment). Are you care about minuses?

veky on July 10, 2014, 8:47 a.m.
<blockquote> I got email that you arguing with yourself </blockquote> That's exactly the problem here. Aiadan's responses consist only of minuses. I don't think it's exactly a productive discussion. <blockquote> Are you care about minuses? </blockquote> Of course. Not directly, but they indicate somebody misunderstood something. Maybe that's me, but minuses don't help me much to understand what's the problem. As far as I can see, Aiadan's main objection is that this solution is in the wrong category, not that it's somehow bad solution. But I fail to see the creativity here. :-)

sheikhmango on July 23, 2014, 10:44 p.m.
Wow this is so good, I have a lot to learn.

veky on July 23, 2014, 11:19 p.m.
True. :-)

sloppyshenanigans on Aug. 14, 2014, 11:03 p.m.
How does this code take into consideration a break in the string for digits? I'm curious as to how I could shorten my code into something like this... This is how I solved it: <pre class='brush: python'>def checkio(words): wordSplit = words.split() count = 0 for i in wordSplit: if i.isalpha(): count += 1 elif i.isdigit() and count < 3: count = 0 if count >= 3: return True else: return False </pre>

veky on Aug. 16, 2014, 8:52 a.m.
That's really too complicated. For one trivial thing, last 4 lines can be just <pre class='brush: python'>return count >= 3 </pre> But more importantly, you can return True whenever you get to count == 3, you don't have to go through the rest of words. My code has word.isalpha() multiplied by count+1. What does it do, is explained in the rest of the comments above. If you still find it unclear, feel free to ask.

catalanomarko on July 1, 2021, 3 p.m.
I try what do u say, but isn't works: <pre class='brush: python'>def checkio(words): count = 0 for word in words.split(): if word.isalpha(): count+=1 else: count=0 return count >= 3 </pre> Your result:False one two 3 four five six 7 eight 9 ten eleven 12 Right result:True

veky on July 1, 2021, 8:29 p.m.
You can't "try what do I say" in one piece of code, since I said two things above. :-D First thing I said was that you can take the above solution, and just replace the last four lines with `return count >= 3`. It works, but you didn't do that. Second thing I said was that you don't need to wait until all the words are processed to return, you can return as soon as you get to 3 with your counter. It also works, but you didn't do that either. :-] Which of those two things did you try, and do you see where you went wrong?

theholy7 on Aug. 26, 2014, 9:17 p.m.
Fantastic answer! I'm amazed, I would never have thought about it!

veky on Aug. 27, 2014, 3:21 p.m.
Hm... besides multiplication with a bool, what else is so ingenious here? :-)

theholy7 on Aug. 27, 2014, 6:21 p.m.
A lot of solutions I see around here are only a couple of lines of code. I always code in a trivial way and I completely miss the simplicity that python can bring to the table. I lack knowledge and experience to be succinct and simple. Like your answer. And that's why I find it so fantastic!

veky on Aug. 28, 2014, 5:05 a.m.
Ah, yes. Well, if you publish your solutions, I'd be able to help you more directly.

JosephBernard on Nov. 3, 2014, 10:18 a.m.
I seem to have the same problem as coming from c++ I go back to the fundamentals and since I don't know all these nifty predefined functions I end up writing them myself. Looking at my solution you could almost directly plug it into a c++ compiler and run it, lol.

veky on Nov. 3, 2014, 10:31 a.m.
Yes, it's a known phenomenon. When you speak language A very well, and then try to learn B, at the start you're still speaking A, even if words come from B. Your mental pathways are still unaccustomed to phrases and expressions of B. It's nothing to be ashamed of, you just need to be aware of it and willing to change it. Read a lot of other's code, and rewrite your code whenever you see something you find generally useful. In about a year, you'll be bilingual. ;-)

theholy7 on Aug. 28, 2014, 10:59 a.m.
I didn't see the publish button ahah. I'll start publishing them from now on! :) Thanks for the tip.

smjefferson602 on Sept. 30, 2014, 1:24 a.m.
There's that Iverson you told me about the other day!

veky on Sept. 30, 2014, 5:02 a.m.
Yes, and much more. See <a href="http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/">http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/</a>#comment-10587 above. :-)

e.roboda on Dec. 10, 2014, 6:07 p.m.
Cool !

veky on Dec. 10, 2014, 6:32 p.m.
Thanks. :-)

sabareeshk on Dec. 11, 2014, 10:07 a.m.
cool

SrdjanTubin on Dec. 30, 2014, 2:45 p.m.
Again, very nice solution, but it's not generic, it would fail if the test case was like "bla bla bla bla 444 bla", and again - I would say it's rather creative than clear. Maybe I'm wrong, correct me otherwise, and take no offence please, we're all here to learn :)

veky on Dec. 30, 2014, 6:20 p.m.
I'm not sure I follow you. What do you think "bla bla bla bla 444 bla" should return, and what do you think my code returns? (My understanding is that it should return True, and it does indeed return True.) I never take offence discussing code only. It's when someone tries to discuss my character, then I explode. ;-D

SrdjanTubin on Dec. 31, 2014, 7:55 a.m.
Sorry...in all my haste, i didn't notice "for word in words.split()", so you actually do check all the words in the input string (wrods). Discussing anyones character in a python course is meaningless...to me at least :)

veky on Dec. 31, 2014, 9:02 a.m.
Ok, now that we've cleared that up, let's continue that other discussion... it is far more interesting. ;-)

jakobEatsCake on Jan. 16, 2015, midnight
why dont you need to worry about proper white space in the if and else statements for the returns after the :? This is interesting

veky on Jan. 16, 2015, 10:15 a.m.
<a href="https://docs.python.org/3/reference/compound_stmts.html:">https://docs.python.org/3/reference/compound_stmts.html:</a> A suite can be one or more semicolon-separated simple statements on the same line as the header, following the header’s colon, or it can be one or more indented statements on subsequent lines. Many people think Python allows only second possibility. ;-) Which is probably fine since only that way you can have nested compounds, and avoid semicolons when having multistatement suite, and all in all have a nicer code, but for simple `return`s (or other "abandon this line" commands, such as `break`, `continue` or `raise`) this one is fine enough. Even PEP8, which is a much stricter coding standard, okays it.

jakobEatsCake on Jan. 17, 2015, 1:32 a.m.
cool, thanks @veky

python003 on Jan. 21, 2015, 9:06 p.m.
Great solution! I can learn so much from you. Thanks so much.

InhumanRampage on Jan. 29, 2015, 3:42 p.m.
I did quite the same but a bit longer :) That 'for else' troubles me a bit, it's weird. oO

veky on Jan. 29, 2015, 5:14 p.m.
You're not alone. Even The Mighty Ned had his problems with those: <a href="http://nedbatchelder.com/blog/201110/forelse.html">http://nedbatchelder.com/blog/201110/forelse.html</a>

InhumanRampage on Jan. 30, 2015, 10:03 a.m.
Thanks ! <blockquote> It seems like "else" should instead be "and also" </blockquote> That definition of Ned fits a lot better to my understanding :) I read further far, and it's very useful ! (actually I was kinda looking for this :D)

veky on Jan. 30, 2015, 11:05 a.m.
<blockquote> actually I was kinda looking for this </blockquote> I have a weird gift, I usually give people what they were looking for, even if they don't know it in the first place. ;-)

lena-py on Aug. 9, 2015, 4:53 a.m.
<blockquote> You’re not alone. Even The Mighty Ned had his problems with those: <a href="http://nedbatchelder.com/blog/201110/forelse.html">http://nedbatchelder.com/blog/201110/forelse.html</a> </blockquote> As a newbie, I didn't find for / else to be confusing at all. The arguments for avoiding it posted in the comments of the link above are not persuasive; however, <a href="https://twitpic.com/4a52sh">this</a> was.

veky on Aug. 9, 2015, 7:04 a.m.
Interesting. Of course, it's hard to draw conclusions when we don't know how representative that sample was. But there are a few things to consider independently of that. First, there is a nice saying in some style guides for Python, <blockquote> Always assume person reading the code knows Python better than you. </blockquote> In other words, don't be afraid to use full expressiveness of Python, as long as you use it idiomatically. Python is usually so readable it can get away with it. (Just imagine such a rule in C++ for example.:)) You can usually guess about right what a construct does even without knowing it formally. Yes, there are a few warts (and for/else might be one of them), but they are nicely documented and you can expect experienced programmers to have encountered them already. More important point: the person writing the poll either doesn't know much Python, or was intentionally trying to confuse people (or both:). Focusing on `break` is just wrong. There could be many ways of abandoning the loop: return, raise,... and even if the loop wasn't abandoned, inner infinite loop might prevent the loop check. Now imagine if one of answers was (objectively more correct): <ul> <li>the `else` suite executes if `expr` was found to be false</li> </ul> How many people do you think would chose this? :-) Analogy with `if` is startling when expressed this way. Tangential point: notice that here we're talking about for...else, not while...else. Here the idiomatic usage is very important. Almost all for...else usages are of the approximate form <pre class='brush: python'>for item in seq: if cond: do something with item satisfying cond break else: what to do if no item in seq satisfies cond </pre> And here I think it's quite obvious how `else` interacts with inner `if`, even if the correct mechanism is not so easy to understand.

grutte_pier on Feb. 24, 2015, 7:54 a.m.
Hm, somewhere in the back of my mind I knew of this for-else construct. But in this particular case I don't think it does anything, does it? Or in other words: if you would remove the else:, and just type return False as the last line, nothing would change, would it?

veky on Feb. 24, 2015, 10:08 a.m.
That's true. But obviously many people learned something reading that "else", so it has educational value. And it's not completely without point: look at the <a href="http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-15168">pseudocode explanation</a> above, and note the "else" word that fits quite naturally in it.

laharah on April 3, 2015, 1:35 a.m.
While I agree that it's kosher to use a bool in this way, I do not think it qualifies as "clear". If someone didn't know about that property of bool this code would appear to be gibberish.

veky on April 3, 2015, 4:46 a.m.
If <a href="http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-27448">someone didn't know</a> that else could be used with for, the code would appear to be wrongly indented. If <a href="http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-23430">someone didn't know</a> you could inline single-statement suites, the code would appear to be lacking newlines. If <a href="http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-22548">someone didn't notice</a> the loop, the code would appear to check only a single run of words. If someone didn't know this was written in Python, the code would appear to be nonsense. And so on. Of course the point here is to learn Python, not to solve the tasks in the most boring way. For that you have Java. :-D

laharah on April 3, 2015, 10:30 p.m.
I guess what I'm trying to say is that this code is concise rather than clear. It is certainly very pythonic, that's not my issue at all. I thought that the point of the "clear" tag was so that beginners to python, or people who struggled with their solution, would be able to understand what's happening, particularly since this is a "beginner" challenge. If a beginner is trying to learn how to count occurrences in a sequence you wouldn't teach them to implement a `defaultdict(int)` or a `collections.counter`, you'd want them to understand the basic underlying idioms first. You'd do: `if i not in d:...` or: `d[i] = d.get(i, 0) + 1` If they can't do one of the above they'll be helpless on any more complex problem involving dictionaries. For-else is a pretty good idiom, and think people should know it, but it's arcane even to some seasoned python devs (despite my efforts, not everyone has read Knuth). Maybe we should try to get checkio a "concise" or "pythonic" tag. It would certainly be more a more entertaining and productive tag, while not tripping up the people who actually need to see a "clear" solution.

veky on April 4, 2015, 5:32 a.m.
I'm glad people still read Knuth. :-) [Though, if I remember correctly, he advocated a much more general approach, roughly equivalent to Python's exceptions. But if you're just pointing out his dislike of boolean flags, then you're correct.:] Now let's go to more serious stuff. The first big problem with your worldview is that the (Python) beginners are very diverse. I know, I've taught them for a long time. And still they manage to surprise me regarding what seems obvious vs what seems impenetrable to them. For example, those who have some kind of C foreknowledge do have problems with lambdas, but not syntactically - they have more problems with the idea of nested functions. On the other hand, those who've grown up on BASIC V2 and DEF FNA(X), grasp lambdas much more easily (despite the ironic fact that the command is named DEF:). Kids in sixth grade, who've just encountered (and liked) the idea of a mathematical function, also find lambdas more intuitive. [Also, list/set/dict comprehensions. Despite seeming "obvious" to Python veterans, it turns out very few beginners really get them right. Especially when iterations are nested, they are very hard for many.] That's why I get an allergic reaction every time someone says "beginners don't get lambdas" (or some other feature). Not all of them. Not everyone goes to Python from C. The world is already too biased towards that learning path. I'm trying to use my enormous reputation here towards rectifying that bias. Second problem: you might think that disassembling some step into more "elementary" steps (quotes because there is really no universal criteria for what "elementary" means) is always a win when explaining, but experience shows it's not the case. It's a tradeoff, same as with everything else. Why exactly is "i not in d" elementary step? And second line will almost always (trust me, I've been there) get you a question "why can't I simply write `d.get(i, 0) += 1`?" :-D And then you have big problems, even if you go down the "well, there's setdefault"-path. :-P ["Underlying operations" are a lie. Why don't you tell them about hashing and buckets and key immutability and other wonderful "underlying operations" stuff when explaining dict? Of course, because they are not relevant. Completely the same, general dict is not relevant to the idea of tallying up a sequence.] While we're at it, why is dict more elementary than Counter? It is certainly more general, in the usual usage. The fact that Counter is implemented by inheriting from dict doesn't really impress beginners at that stage. They just see something more immediately useful for their task at hand. Why would you want to hide it from them, just because in your neatly arranged view of a person who already knows all this stuff, dict "comes before" Counter?

drobotnik on April 10, 2015, 10:37 a.m.
Why did you put the else in? Why not: def checkio(words): <pre class='brush: python'>succ = 0 for word in words.split(): succ = (succ + 1)*word.isalpha() if succ == 3: return True return False </pre>

veky on April 10, 2015, 10:56 a.m.
<a href="http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-25807">This is why - kinda.</a>

drobotnik on April 10, 2015, 11:07 a.m.
I guess I missed that else can be used with for! My eyes just skipped over that part. Cheers!

Vedeshkin on June 2, 2015, 9:48 a.m.
veky you just optain a new fan ;-) Thanks to you, I learned a trick with multiplying types ;) Thanks!

veky on June 2, 2015, 5:58 p.m.
I surely hope you're not multiplying types. :-) <pre class='brush: python'>>>> int * str TypeError: unsupported operand type(s) for *: 'type' and 'type' </pre> Though even that could be defined with metaclasses*, it's not really supported. (Cartesian product is spelled differently.:) You meant "multiplying objects of different types". And yes, bool is just one of them. Try `"wet"*3` or `5 * (3,8,1)`. :-) <hr /> *Since someone will surely ask me how: <pre class='brush: python'>>>> class multified(type): __mul__ = __rmul__ = lambda *a: "{.__name__} * {.__name__}".format(*a) >>> class int(int, metaclass=multified): pass >>> int * str 'int * str' </pre>

Shoneson on June 14, 2015, 1:17 p.m.
Having returned to programming after 10+ years (and choosing Python for that), I try to find a solution on my own. Being a Python beginner I have a beginner's approach and way of thinking. I am aware that I have a lot to learn until I start 'breathing' Python. But, when I finish the task on my own, I always look forward to 'unwrap' one of your solutions, and it seems like looking at Mt. Everest from a base camp. Out there in the distance, saying 'I am waiting for you...' And thinking: Not this time, but some day definitely. After all, people have been known to conquer it. It is not the destination, but the journey that matters.

veky on June 14, 2015, 1:49 p.m.
Thank you for the kind words. And I congratulate you for choosing the best language in the universe for your return. ;-) Of course, my main message is never that you have to breath Python to be successful, just that the world is much more beautiful, and things make much more sense, when you do. :-) For completeness, I must say that even I am not quite at the summit. There is so much to learn, so many ideas to incorporate. Even weirder, the summit is moving. Just two weeks ago, Py3.5 gave us few things that I would never have thought a Python would have. I have to rewrap my head around these, to find how to fit them into the overall philosophy of the language.

fh2game86 on July 25, 2015, 2:43 p.m.
hahahah you put it all together to look small

veky on July 25, 2015, 6:25 p.m.
I must say I don't understand you. You mean returns on the same line as ':'? And, independently, that's reason enough for -2 for you??

Czarinov on July 30, 2015, 7:36 p.m.
Nice! isalpha() was what I was looking for. Instead I had to create ck='0 1...9 ' and check if ws[i-1:i] in ck But now I know :P. Anyway, what's about some people not accepting True == 1... Didn't they ever learn logical multiplication or addition? That's basic math :D. Thanks for more knowledge!

veky on July 30, 2015, 9:02 p.m.
`ck` is in fact `string.digits`, right? Though I don't really know what the spaces are doing there... And `ws[i-1:i]` is just ws[i-1] if ws is a str. They are probably ok with logical multiplication (`and`), but this is half-logical multiplication (bool times whatever), and not many people realize that Python's `and` works that way too. <pre class='brush: python'>True and whatever ~~~> whatever (not necessarily True) False and whatever ~~~> False True * 5 ~~~> 5 (not 1) False * 5 ~~~> 0 </pre>

Czarinov on July 31, 2015, 9:03 p.m.
I am still learning and my solution was a bit tryhard. I already changed it to better one, so I can't copy what original was. Original idea was something along: <pre class='brush: python'>ck='1 2 3 ... 0 ' count=0 for i in range(1,len(ws)): if (ws[i]=' ') * (not ws[i-1:i] in ck): count+=1 ... </pre> But it doesn't make any sense now why I used space ' ' in there... no idea, was late evenin and I was typing tired. xD

veky on Aug. 1, 2015, 11:44 p.m.
<pre class='brush: python'>if cond: count += 1 ~~~> count += cond </pre> Or just use sum. ;-)

Czarinov on Aug. 2, 2015, 8:33 a.m.
yes, bool to win! :D

veky on Aug. 2, 2015, 4:28 p.m.
:-) Iverson bracket is a powerful concept.

natter1 on April 3, 2021, 12:39 p.m.
Do you have an idea, why so many sources out there explain 'and' (same for 'or') wrong? I even stumbled at explanations that where correct but still claimed wrongly "... returns True ..." in the beginning.

veky on April 3, 2021, 1:33 p.m.
Because Python is carefully designed. ;-P It might seem like a contradiction, so let me explain. It's extremely hard to design a language that will be attractive to both beginners and experts. But it isn't impossible: since many of areas of interest don't overlap, you can have a language where both newbies and pros have something to enjoy. But on the other hand, it's practically impossible to document such a language: beginners see only things that matter to them, not being aware of layers upon layers of features waiting to be used; and experts find it really hard to figure out which of tens of technical details are what's important to people who see the feature for the first time. To someone who didn't even think to try to write `5 and {8, 2}` into the REPL, it sure seems that `and` does the Boolean conjunction and nothing else -- because that's exactly the projection of its behavior to {True, False}. Math is the same way: you might think that "power" means just repeated multiplication, and you won't know what is e^pi (let alone e^(pi*i):), but you will know 3^4=81, and you might think that's all there is to it. There are many "math documentations" where that's exactly the official point taken. You'll never find out about <a href="https://www.youtube.com/watch?v=O85OWBJ2ayo">this</a> unless you get out of your comfort zone. Let me shock you a bit more: almost everything you see written about Python online is wrong. Even the official Python documents sometimes lie in order to be easier to understand. But the right question is not to ask "is this wrong", but "in which ways is this wrong" -- that is, "can this be generalized in a way that's more useful to me". In astonishingly many cases, the answer is yes. :-)

natter1 on April 3, 2021, 2 p.m.
Thanks for the reply. Honestly, while I get the general idea, I really don't see the point, why to promise people a function (or a statement? or result of an expression?) returns either True or False, when it in many cases gives a completly different result. And I stumbled to understand some solutions here in the beginner area due to that (like log10(x or 1)). Its not that I want to know about all edge cases, but why give explanations, that prevent people to understand simple code? On the other hand, I was also the guy that argued with my teacher back in the days, that water cannot be truly incompressible (because that also would create so many issues) ^^. Its ok to use a simplified model, but you have to name it as such. Guess I have to be more careful with my python sources :) The links you give in the comments look promising.

veky on April 3, 2021, 5:50 p.m.
You make it sound as if someone is doing that intentionally. While it does happen sometimes, in many cases people write that because they really don't know better. If someone asked you what is exponentiation, you'd probably use repeated multiplication as the basis for your explanation (unless you're a professional mathematician). Does it make you a bad person? No, it just makes you someone who doesn't realise the full depth of what they're talking about. :-)

natter1 on April 4, 2021, 12:34 a.m.
I'm not sure what is worse - if people who are educating others do it to make it easier to understand, or if those people really don't now better (and arn't corrected for years). When I google python 'or' I get <a href="https://www.w3schools.com/python/python_operators.asp">w3schools</a>. They look like they should know what they are talking about to me. But for 'or' they state (and I'm pretty sure thats on their site for years already): "Returns True if one of the statements is true" instead of e.g.: " x or y: x, if it evaluates to true, otherwise y" Of course it happens even to the greatest to be "someone who doesn’t realise the full depth of what they’re talking about". But surly there are people out there, who could point out such (in this case) simple mistakes. Btw. thanks for the video "How (and why) to raise e to the power of a matrix" - I wish our profs. would have used such cool animations during their lectures :)

veky on April 4, 2021, 4:39 a.m.
Ah, w3schools. A typical example of people who really don't give a damn about quality of their materials. Please read <a href="https://www.quora.com/What-is-wrong-with-W3Schools-that-it-is-often-referred-to-as-a-bad-resource-for-learning">this</a> for a full view. Matrix exp: I just wanted to show you that a well-designed system is always richer than anything you can fixate in words. And there's no reason to hate it.

lena-py on Aug. 9, 2015, 4:29 a.m.
Simple question: for what does "succ" stand? success?

veky on Aug. 9, 2015, 5:08 a.m.
Hm, no. "successive" was on my mind (as in, "successive word count"). Just shows how abbreviations are almost never worth it, even if they are used four times. :-|

widzard on Aug. 27, 2015, 2:38 a.m.
I just know about " number * bool ", great , and thank you !

veky on Aug. 27, 2015, 4:21 a.m.
bool is number in Python. <pre class='brush: python'>>>> from numbers import Number >>> issubclass(bool, Number) True </pre> ;-)

arekrob on Aug. 30, 2015, 10:34 a.m.
I agree with some of the above comments that multiplying by a boolean is "hackish". When I read the code, in my head I have to convert that multiplication to if ("if isalpha then keep the value, else set it to zero"). Maybe for such a small piece of code it's ok, but if it was part of a production code of some real application, I think it would not pass code review.

veky on Aug. 30, 2015, 11:20 a.m.
Depends on who is doing the review, of course. :-) A question for you: when you see `cond1 & cond2`, do you also 'have to convert that conjunction to if ("if cond1 then cond2 else False")'? This is completely analogous. You can split cases, but some operations really are elementary.

MhdTahawi on Sept. 15, 2015, 9:06 a.m.
Thanks, I learnt something new today :)

veky on Sept. 15, 2015, 1:14 p.m.
That's great. :-)

alexanderkopke on Oct. 5, 2015, 10:54 a.m.
simple, educational, brilliant! Thanks!

veky on Oct. 5, 2015, 11:11 a.m.
You're welcome. :-)

totee19 on Oct. 21, 2015, 8:27 p.m.
Your solution is excellent but not clear cause I hate this syntax: <pre class='brush: python'> if succ == 3: return True else: return False </pre>

veky on Oct. 22, 2015, 12:19 a.m.
You hating some syntax surely doesn't mean the solution is not clear. And I have a feeling that your hate stems from wrong pairing. What you quoted is not a sub-unit of the code above, since `else` is not paired with that `if`, but with `for`. It's as if someone wrote <pre class='brush: python'>result = [] for item in sequence: with contextmanager(item): result.append(item) return result </pre> and you wrote that you hate the syntax <pre class='brush: python'> result.append(item) return result </pre> "That syntax" is, in a certain sense, not a part of what I wrote, in a similar way that `a) +` is not a part of expression `abs(a) + b`. :-) You hating `a) +` shouldn't be an argument for finding `abs(a) + b` unclear.

totee19 on Oct. 22, 2015, 2:24 p.m.
I understand your notice. For me when a solution is "clear" it's when I judge/can understand easily. I would have prefered something like that: <pre class='brush: python'>def checkio(words): succ = 0 for word in words.split(): succ = (succ + 1)*word.isalpha() if succ == 3: return True else: return False </pre>

veky on Oct. 23, 2015, 7:20 a.m.
Ah, so I completely misunderstood your objection (which is in part because it was so original... there are many things people didn't like about my solution, but you're the first one about this:). You meant the compact writing of one-line suites, after the colon without introducing another line. This is mostly an aesthetic concern, and as such inherently subjective. For example, I find it ugly that in your version `else:` is dedented by more than its length from the previous line, producing an optical illusion that there is an empty line between them. Also, I think that since `if` and `else` aren't on consecutive lines, it's easier to think that your example was accidentally misindented. But there're other concerns here on CiO. As you're probably aware, I'm trying to teach people here various things about Python. And <a href="http://www.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-23430">at least one person</a> learned something new. And it is not something obscure: PEP8 speaks about it, and even okays it in some cases when a clause "stands for itself". It's also important to accentuate the fact that those are not only one-line suites, those are `return`s. The usual objection---that someone might add some code later into the wrong block---thus doesn't hold, since it would be unreachable code anyway.

Evgeny_Kuizhev on Nov. 12, 2015, 10:36 p.m.
Very well

ducthinhdt on Nov. 14, 2015, 12:28 p.m.
after dizzy my head, i learned something new. Cheers.

veky on Nov. 14, 2015, 1:22 p.m.
As usual. ;-)

nickb on Nov. 18, 2015, 6:05 p.m.
This is nice!

veky on Nov. 18, 2015, 6:22 p.m.
Thanks.

dalluca on Feb. 28, 2016, 7:25 p.m.
So, in facts, <pre class='brush: python'>succ = (succ + 1)*word.isalpha </pre> is the same as <pre class='brush: python'>if word.isalpha(): succ += 1 </pre> But, are there computational differencies? I think the second solution is computationally better because if word.isalpha() is false you avoid to evaluate succ+1; and I think it is clearer too; let's tell it is less matematic. What do you think about?

DahliaSR on May 8, 2016, 4:40 p.m.
That's a nice solution, multiplying int and bool is handy here. But your solution iterates through words.split() completely, even thoungh there might be no need for it.

veky on May 9, 2016, 6:32 a.m.
No, it doesn't. Line 5 stops function (and loop) execution as soon as the answer is known.

DahliaSR on May 9, 2016, 5:27 p.m.
That's not what I had in mind. Let's assume the string passed to the function is: <pre class='brush: python'>"Hi, my name 123 is DahliaSR" </pre> As soon as the loop checked "123", succ can not become 3, as there are only 2 indexes left to check. I took care ofthis in my solution , but it turned out to be ineffective.

veky on May 9, 2016, 9:33 p.m.
Well, of course that's ineffective. You're investing another check inside the loop, with a lot of arithmetic, that is checked every time through the loop, only to spare yourself of at most 2 checks - asymptotically negligible. Doesn't seem like a good ROI. :-)

ljy95135 on May 24, 2016, 8:58 a.m.
I like it! Every time I see the answer on the top, it shocks me again for how short it can be.

veky on May 24, 2016, 10:07 a.m.
It is short because the task is simple. Ideal programming language should have solutions of approximately the same length (entropy) as the task text. And Python is really close to that. :-)

inkimiaei on June 6, 2016, 4:36 p.m.
Great Solution :)

veky on June 6, 2016, 4:41 p.m.
Thanks.

jiangwei0515 on Aug. 3, 2016, 12:18 a.m.
Thanks for the solution. I tried debug the code with the output as follows. It seems that having the else in line 7 will make the code always return a false? am I missing anything? checkio("Hi my name 123 is DahliaSR") <pre class='brush: python'>1. def checkio(words): 1. succ = 0 1. for word in words.split(): 1. succ = (succ + 1)*word.isalpha() 1. print ("Word:{}, succ:{}").format(word,succ) 1. if succ == 3: print "==> true" 1. else: print "===>false" 1. 1. Word:Hi, succ:1 1. Word:my, succ:2 1. Word:name, succ:3 1. ==> true 1. Word:123, succ:0 1. Word:is, succ:1 1. Word:DahliaSR, succ:2 1. ===>false </pre>

veky on Aug. 3, 2016, 5:55 p.m.
You are missing a lot of things. :-) First, please try to use Py3. Your life will be much easier in the long run. Second, what is your checkio function supposed to do? If you're talking about this task, then checkio should return the bool value, not print it. And third, obviously, even if `else` did what you want there, you wouldn't want to print "true" every time succ became 3, right? Especially since you don't print "true" when it's greater than 3. :-] So, the idea is obviously to break the loop once you print "true". And then else will work as it's intended.

jiangwei0515 on Aug. 3, 2016, 9:25 p.m.
Thanks Veky. I'd like to understand how this for-else work hence I debug it by printing the steps. I was hoping the output to be true when I run checkio(“Hi my name 123 is DahliaSR”), however it returns false as the last line of the output indicated. Am I missing something here?

veky on Aug. 4, 2016, 2:50 a.m.
Sorry, but it doesn't return anything. <pre class='brush: python'>>>> returned_value = checkio('Hi my name 123 is DahliaSR') ...... a bunch of prints ...... >>> print(returned_value) None </pre> As I said, returning is not the same as printing. Look. Suppose your task is to write a function to calculate a Collatz transformation of a natural number n (divide n by 2 if you can, else multiply by 3 and add 1). Of course, the solution is <pre class='brush: python'>def Collatz(n): half, odd = divmod(n, 2) if not odd: return half return 3*n + 1 </pre> Now you can use that function as a building block for more complicated functions. For example, you could calculate the length of the Collatz sequence starting with n (the number of Collatz transformations n takes to reach 1). <pre class='brush: python'>def lenCollatz(n): steps = 0 while n > 1: steps += 1 n = Collatz(n) return steps </pre> Every one of them might have its result printed, either by using `print` function in the code, or just by typing its call into the REPL. <pre class='brush: python'>>>> Collatz(5) 16 >>> lenCollatz(9) 19 </pre> But note that - and this is crucial - neither of these functions uses `print` at all. `print` is an external wrapper, either implicit (in REPL) or explicit (in scripts), but it prevents the very useful combination of functions as I wrote above. Suppose you wrote <pre class='brush: python'>def bad_Collatz(n): half, odd = divmod(n, 2) if not odd: print(half) print(3*n + 1) </pre> Now REPL still "works" <pre class='brush: python'>>>> bad_Collatz(5) 16 </pre> but it behaves weird in some cases <pre class='brush: python'>>>> bad_Collatz(6) 3 19 </pre> and much more importantly, it's practically useless for building further functions like lenCollatz. `print` does a bunch of transformations to your object, some of which are irreversible, making it useless for almost everything further. It first converts your object to a string. Then it encodes that string into a bunch of bytes according to your system encoding. Then it puts those bytes into a special area of memory ("buffer") of your standard output. Then it flushes that buffer (usually), converting them to a bunch of low-level system calls. In the end, you have some squiggles on your screen, but you don't have your original object anymore. Visualization is not the data. Selfie is not the person. Map is not the territory. <hr /> Ok, you might say, but I just used print for debugging, not for returning. First, there are much more useful tools for debugging (logging and pdb come to mind - if you want to know how to use them, say so:). And second, more importantly, that doesn't change the fact that you didn't use anything for returning, relying on implicit "return None" at the end of the function. Using print for debugging is (while clumsy) understandable, but you still have to return something from your function. If you want to see what a function returned, just call it from the REPL. :-)

Booplesnoot on June 9, 2020, 3:05 p.m.
Hi there, I would to know how to use logging and pdb for debugging instead of using print :)

veky on June 9, 2020, 6:58 p.m.
logging: <pre class='brush: python'>from logging import * getLogger().setLevel('DEBUG') def checkio(words): succ = 0 for word in words.split(): succ = (succ + 1)*word.isalpha() debug('Word=%r, succ=%d', word, succ) if succ == 3: return True else: return False info('==> %r', checkio('Hi my name 123 is DahliaSR')) </pre> new style (in the meantime Python has relented and made print much easier for debugging): <pre class='brush: python'>def checkio(words): succ = 0 for word in words.split(): succ = (succ + 1)*word.isalpha() print(f'{word=}, {succ=}') if succ == 3: return True else: return False print(f"{checkio('Hi my name 123 is DahliaSR') = }") </pre> <hr /> pdb: <pre class='brush: python'>def checkio(words): succ = 0 for word in words.split(): succ = (succ + 1)*word.isalpha() if succ == 3: return True else: return False breakpoint() print(checkio('Hi my name 123 is DahliaSR')) </pre> After starting such a program: <pre class='brush: python'>> forelsedebug.py(9)<module>() -> print(checkio('Hi my name 123 is DahliaSR')) (Pdb) s --Call-- > forelsedebug.py(1)checkio() -> def checkio(words): (Pdb) n > forelsedebug.py(2)checkio() -> succ = 0 (Pdb) n > forelsedebug.py(3)checkio() -> for word in words.split(): (Pdb) n > forelsedebug.py(4)checkio() -> succ = (succ + 1)*word.isalpha() (Pdb) n > forelsedebug.py(5)checkio() -> if succ == 3: return True (Pdb) word, succ ('Hi', 1) (Pdb) n > forelsedebug.py(3)checkio() -> for word in words.split(): (Pdb) n > forelsedebug.py(4)checkio() -> succ = (succ + 1)*word.isalpha() (Pdb) n > forelsedebug.py(5)checkio() -> if succ == 3: return True (Pdb) n > forelsedebug.py(3)checkio() -> for word in words.split(): (Pdb) n > forelsedebug.py(4)checkio() -> succ = (succ + 1)*word.isalpha() (Pdb) n > forelsedebug.py(5)checkio() -> if succ == 3: return True (Pdb) word, succ ('name', 3) (Pdb) n --Return-- > forelsedebug.py(5)checkio()->True -> if succ == 3: return True (Pdb) n True --Return-- > forelsedebug.py(9)<module>()->None -> print(checkio('Hi my name 123 is DahliaSR')) (Pdb) c </pre> About pdb, I've tutored the IDLE master, Terry J. Reedy, <a href="https://bugs.python.org/issue27821#msg273535">here</a> (read the rest of the thread). <hr /> Also, for sure, <a href="http://www.pythontutor.com/visualize.html">http://www.pythontutor.com/visualize.html</a>#mode=edit should be mentioned. Just paste your code and click Visualize Execution (you might also select "render all objects on the heap" for a more faithful but clumsier representation). If there's anything more you'd like to know, feel free to ask.

laerus on Sept. 25, 2016, 3 p.m.
Readability counts. Explicit is better than implicit. Simple is better than complex. import this

veky on Sept. 26, 2016, 7:52 a.m.
... your point being? Of course I find this readable, explicit and simple. If you don't, please give concrete arguments.

glenuk on Oct. 17, 2016, 11:43 p.m.
Interesting, never thought of using a bool like this.

veky on Oct. 18, 2016, 7:15 a.m.
That's exactly why I wrote it this way. :-)

satoshi1117 on Oct. 26, 2016, 8:08 a.m.
your answers always clear and beautiful.

veky on Oct. 26, 2016, 12:12 p.m.
Tell it to <a href="https://py.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-45725">this guy</a>. :-D

ermichin158 on Nov. 20, 2016, 10:25 a.m.
Nice solution. Let's contvert all booleans to int(or ints to booleans... idk :))!

veky on Nov. 20, 2016, 1:01 p.m.
I don't understand you. All booleans are ints. What do you mean by "converting" them? <pre class='brush: python'>>>> isinstance(True, int) and isinstance(False, int) True </pre>

ermichin158 on Nov. 20, 2016, 1:04 p.m.
Yes, i know. I used them as int too in my language. And... idk what to say... I dont understand myself too, sry :~]

veky on Nov. 20, 2016, 1:12 p.m.
<blockquote> used them as int too in my language. </blockquote> Your language? :-)

olegpython1994 on Jan. 22, 2017, 11:03 a.m.
Very pretty solution!)))

inkplay_ on March 22, 2017, 5 p.m.
Damn that's elegant my friend. I'm going to steal this. Using a bool as 1s and 0s and automatically reset the counter to 0 when it multiples a false statement, that's smart.

EmmaJiro on March 23, 2017, 9:01 p.m.
Thanks ! I like it !

veky on March 24, 2017, 3:05 p.m.
You're welcome. :-)

marlonxteban on April 2, 2017, 3:08 a.m.
LOL Veky... I forgot the simple bool * int hahaha... great!

veky on April 2, 2017, 9:51 p.m.
That's why I write these solutions... to remind you. :-D

AndrewSaltz on June 4, 2017, 10:07 a.m.
This is fun. Thank you

veky on June 6, 2017, 2:03 a.m.
You're welcome. :-)

leandroribeiro on July 4, 2017, 10:53 a.m.
Do you really need that final else statement? Wouldn't 'return False' work just as well?

veky on July 4, 2017, 1:19 p.m.
Of course I don't. But please read other comments (e.g. <a href="https://py.checkio.org/mission/three-words/publications/veky/python-3/for-else/share/cc9483a7db9cabc4f132f0dfcfa37b46/#comment-25797">this one</a>) here for a perspective.

leandroribeiro on July 4, 2017, 2:10 p.m.
Thank you.

erickfis on Oct. 23, 2017, 1:53 p.m.
awesome job!

Cocoolanu on April 15, 2018, 8:58 a.m.
To be a Python expert, I think we definitely should avoid to use else blocks after loops as this behaviour isn't intuitive and can be confusing. I think we should avoid to teach them to use for...else.., especially for beginners.

veky on April 15, 2018, 3:03 p.m.
What is your definition of "expert"? For me, it's somebody who uses the full potential of their tools. Also, I'm sure you're aware that many other features of Python "aren't intuitive and can be confusing" (at least according to some people). I'm quite convinced that if we removed all of them, the language would feel too restricted even for you. Just an example: your solution has multiple exits, which can be confusing, and uses +=, whose behaviour isn't intuitive. :-) Of course, I don't mean anything wrong about your solution, I'm just telling you your criteria are incredibly broad.

Leetroch on June 18, 2018, 3:34 p.m.
Your solution is awesome, but I need to make sure that else：return False can be dispatch out of for loops

veky on June 19, 2018, 11:45 a.m.
Of course they can. It's in the language specification. :-)

hugoboss on Aug. 7, 2018, 3:06 p.m.
BOOLs IS INTs !!! Thanks! It's awesome. And the sixth line is unnecessary.

veky on Aug. 8, 2018, 12:11 p.m.
Bools are ints, yes. :-) Well, I think that when I first solved that task, it didn't accept any true/false objects, only True/False objects as return values. I'm glad they changed it now. :-)

alekseykonotop on Sept. 25, 2018, 3:46 p.m.
Very cool!

t4d3uszk on July 18, 2019, 8:41 a.m.
Very clever. I didn't get the idea of multiplying bool by int but when I see your solution it is so obvious ;-)

veky on July 18, 2019, 9:41 a.m.
Read about the Iverson bracket some time. ;-)

bsquare on Sept. 18, 2019, 1:41 p.m.
Very nice solution, I read alot of the comments here, it was interesting, thanks @veky!

fgfgfz on April 2, 2020, 1:44 a.m.
Nice

fgfgfz on April 2, 2020, 1:45 a.m.
Nice

piter239 on April 21, 2020, 8:12 a.m.
wow)) back to the basics, in an pretty way! I think this is the way it was done in C in early days.

veky on April 21, 2020, 8:25 a.m.
Haha, not even close. :-P Let's just count the obvious high level stuff used here that C doesn't even approach. <ul> <li>direct (object oriented) iteration through things you're interested in, instead of `for(int i=0;i<some_arbitrary_number;++i)\` iteration through indices * \`.split\` method that gives you all the pieces at once, instead of \`strtok\` from \`<string.h>` which must work from the pinhole of static memory</li> <li>`for...else` pattern, which elegantly frees the programmer from maintaining program flow information in boolean flags</li> <li>strings as a primary datatype, without chars as an "even more primitive" -- so things like `.isalpha` work on strings directly, they don't have to be mapped over them as `isalpha` from `<ctype.h>`</li> <li>real (Unicode) characters that form words (and separating whitespace), instead of a 7-bit kludge not enough to write even decent English</li> <li>true booleans (e.g. returned by `.isalpha`), which have a concrete semantics as integers (`True==1` enabling me to multiply with it) instead of "just return something nonzero" from the function in `<ctype.h>`</li> <li>dynamic (executable) function definitions, which enable environments such as these (you can't write CiO in C)</li> <li>not having to declare variables, since they are just names for objects existing independently, instead of boxes of finite size and interpretation</li> <li>not having to declare types, since types reside with those who really should have them: objects, not variables</li> <li>not having to compile and link as a separate step, since everything is available at runtime</li> <li>not having to write braces, since redundancy with indentation (with you should use anyway) often gives mixed messages, and creates visual noise</li> </ul> It feels nice to romanticize the past, but really, it was a much harsher environment than we remember. :-]

piter239 on April 21, 2020, 10:18 a.m.
you are absolutely right!

MrFitz on April 21, 2020, 8:46 p.m.
Wow. You've made a very efficient solution :-) I'm looking back at my code now and thinking how cumbersome it looks! I guess if I had known about .isalpha() then I could have written it more efficiently!

veky on April 21, 2020, 9:27 p.m.
Depends on what you mean by efficient... performance was not my aim. I tried to write an elegant solution, without extra moving parts. :-)

MrFitz on April 21, 2020, 9:42 p.m.
I suppose I meant written in the least code required to do the same job. I like the word elegant to describe a person's coding style. I've just noticed that you posted your solution over 6 years ago and people are still commenting on it! Like posting on a forum, only to later notice that the last person's comment was written years ago. This website is obviously thriving!!

veky on April 21, 2020, 10:25 p.m.
The code doesn't get old... at least not as quickly as posts on various other social networks. :-)

kim.yangjin on May 1, 2020, 6:20 p.m.
I enjoy comparing your solution with mine, even if I get frustrated all the time :)

veky on May 1, 2020, 7:19 p.m.
It's important to become less and less frustrated with time. ;-)

lopeslusca11 on May 16, 2020, 7:52 p.m.
I thought it was great, especially the Boolean multiplication, I hadn't thought about it.

UlanDT on July 20, 2020, 7:16 a.m.
this is amazing!

undead2k on Aug. 1, 2020, 10:23 a.m.
6 years on and still teaching us how to think outside the box with python. Always a pleasure to see your solutions and to read through the comments :D

veky on Aug. 3, 2020, 9:19 a.m.
Yes, CiO is really a social network in my taste. :-D

VincentDog on Oct. 15, 2020, 5:33 p.m.
Smartass mf:)

veky on Oct. 15, 2020, 7:28 p.m.
I'm in the Land of Weird Compliments again. :-]

VincentDog on Oct. 16, 2020, 1:25 p.m.
haha) I'll try to be less weird next time:)

JulesSoulfly on Nov. 27, 2020, 12:11 p.m.
Do you really need 'else' here?

veky on Nov. 27, 2020, 12:36 p.m.
No. :-) Read the above comments.

svalazar on Dec. 23, 2020, 10:25 a.m.
It is a very elegant solution!

shchekalev.alexandr on May 3, 2021, 4:37 p.m.
wow, it's awesome

CIOHarazi on May 4, 2021, 10:43 a.m.
This is really smart. Love it

_Chico__5d82cee4f5234e78a5c65e on May 12, 2021, 11:58 a.m.
Man your solutions are just GREAT!

_Chico__5d82cee4f5234e78a5c65e on May 15, 2021, 6:41 a.m.
It‘s like impossible to like all of your solutions =/

veky on May 15, 2021, 6:42 a.m.
You mean they are so different, there are too many, or there are some of them that are impossible to like all by themselves? :-]

_Chico__5d82cee4f5234e78a5c65e on May 15, 2021, 7:06 a.m.
460 solutions.... If I would like all of your solutions with +5.... then you will just get 2300points ;) but I can’t , because I haven’t solved every task you’ve solved...

veky on May 15, 2021, 7:09 a.m.
<blockquote> you will just get 2300points </blockquote> ... which will be a drop in the bucket, since I have about 5e4 points here. :-P And no matter what you think of them, surely not all of them deserve +5. ;-) [Some of them don't even deserve a +1, but that's a different story.;]

_Chico__5d82cee4f5234e78a5c65e on May 15, 2021, 7:13 a.m.
Yes,...wait...euh no =C

cult_of_memes on Sept. 28, 2021, 10:17 p.m.
One of the few times the top voted solution is actually something I agree with :P

veky on Sept. 29, 2021, 8:11 a.m.
Sorry about that. ;-]

jjrj1203 on Feb. 2, 2022, 5:37 p.m.
Very new to python, but why does "He is 123 man" not make succ == 3 using this code? Does *word.isalpha() reset succ to zero? Sorry if this is a dumb question, but still trying to figure all of this out.

nov.a on April 12, 2022, 9:52 a.m.
I'm impressed by not using the increment operator!! It's more concisely for initializing value of succ.

veky on April 12, 2022, 11:16 a.m.
It's actually for resetting the value of succ. :-)

viktor.chyrkin on July 7, 2022, 7:29 p.m.
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnice

veky on July 9, 2022, 6:52 p.m.
Any reason for that many ns?

viktor.chyrkin on July 11, 2022, 5:43 p.m.
youtubers.

veky on July 11, 2022, 5:45 p.m.
I guess I'm getting too old. :-)

mikaeldovbnia on Oct. 26, 2022, 6:40 p.m.
You can delete the else in the solution because if succ won't reach three it will automatically return False.

veky on Oct. 27, 2022, 1:25 a.m.
... as already said about 7 times in the comments above. :-]

mikaeldovbnia on Oct. 27, 2022, 11:34 p.m.
There are too many comments above, so I didn't see them. Sorry.🙁

Karing on April 17, 2023, 10:07 a.m.
Hello, I find your solution great. A little amelioration: You could leave the last else out, and write the return False after the for-loop. Best regards Karing

veky on April 17, 2023, 10:09 a.m.
I think you should read the other comments. 😏

Karing on April 17, 2023, 1:18 p.m.
sorry I'm only beinnning wit checkio

aniseo on Aug. 20, 2023, 6:23 p.m.
I'm struggling to undestand why we can use isalpha() here (in the whole task, not really only in this solution). I used isdigit() for restarting the count as I assume that, for example, "I'm" is still a word yet it fails isalpha() checks, and I see no mention in the preconditions that words do not contain special characters. Thoughts? What I missed in the task text that suggests the assumption of no "'"s (or ",", ".", "!"...)?

veky on Aug. 21, 2023, 8:34 a.m.
Yes, the precondition is really badly written, but "The input contains words and/or numbers" should mean that when `.split()`, "words" are either `.isalpha()` or `.isdecimal()`.

t3st3rr0rs on Feb. 4, 2024, 10:37 p.m.
it is beautiful solution but I spent minimum 5 minutes to understand how it works

veky on Feb. 5, 2024, 4:17 a.m.
ah, youth these days. In our time, 5 minutes was considered a small time interval. ;)

Strings and Integers

0 %

You ClassRoom Demo is Ready With the new tool CheckiO ClassRoom, you will be able to analyze your students' progress, have your own leaderboard, change missions order on the map, and so on.