filter clear solution for Secret Message by veky - python coding challenges

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

filter solution in Clear category for Secret Message by veky

find_message = lambda text: ''.join(filter(str.isupper, text))

Sept. 3, 2014

Comments:

famaral on Sept. 17, 2014, 7:57 p.m.
This one should be in "Creative" category right?"

veky on Sept. 17, 2014, 10:29 p.m.
Oh no, not another one. :-/ What's creative here? lambda? Or filter? Or .join? Things are used in exactly the way they are supposed to be used (except possibly unbound .isupper). Can it be clearer?

asharkey3 on Sept. 26, 2014, 8:19 p.m.
Could probably be less condescending here hahah. It's "creative" because there are far "simpler" answers for the beginner coder. I myself don't know what lambda does.

veky on Sept. 27, 2014, 5:19 a.m.
<blockquote> Could probably be less condescending here hahah. </blockquote> Probably. I'm sorry about that. But that's kinda my image here. :-D <blockquote> It's "creative" because there are far "simpler" answers for the beginner coder. </blockquote> The fact that you put "simpler" in quotes tells much. You don't really believe that there are absolute criteria for simplicity, especially in this context? Look, I've taught beginners. One of first things you learn about them is that they are very diverse. They like to experiment, and our uniformity of coding standards and PEP8 mean nothing to them. Many of them would solve a task in a way you'd never imagine, convoluted and head-scratching and eye-stabbing. And that's ok, because the primary thing isn't them solving the task, but learning Python. If my solution was somehow "canonical", or "officially supported", of course I wouldn't write it this way. But it is just one of many. And it shows some way to solve the task that might not be immediately obvious, either to a beginner or to experienced Pythonista. By your opinion, only one solution could be in the "Clear" category? I don't think "clear" means that. Yes, there is a line in Zen, "There should be one-- and preferably only one --obvious way to do it.". But "clear" is not the same as "obvious". "Clear" is more like "obvious in retrospective": once you see it, you know what it does, despite maybe not having thought about it in the first place. <blockquote> I myself don't know what lambda does. </blockquote> Ok, imagine this scenario: you have to multiply all the elements in a sequence. Of course, the clearest way to do it is <pre class='brush: python'>product = 1 for element in sequence: product *= element </pre> But is it obvious? Apparently, from what I see during teaching, and here on checkio, it isn't (for everybody). If someone comes from some unnamed programming languages, while loop is more obvious. Now imagine he says: "this is creative because there are far simpler answers for a coder like me. I myself don't know what for loop does". And he's right. He's not trolling (this is important!). Of course, he does know what for loop does in C (oops, I named it:), but that doesn't really help you understand what it does in Python, unless you think on a very high level of "iterating through a container", that many programming beginners don't. For him, Python for loop is something mysterious, and he sticks to what he knows, because while loop, although wordy and convoluted, at least works the same way (at a first glance at least) it works in Java (oops, named it again;). But is this a reason not to use for loop in the above code? NO, for Python's sake. I hope we agree on this. The whiner should just bite the bullet and learn how to use a for loop, and when to use it. It's not hard. Now, the time for learning. If I wrote: <pre class='brush: python'>def find_message(text): return ''.join(filter(str.isupper, text)) </pre> that would be ok with you? But that's almost perfect analogy with the above (for vs while). You insist on a more boilerplatey way of writing something, only because it reminds you of what you know from other languages (defining a function). Right? It doesn't help anybody, not even you in the long run. We're here to learn, not to solve tasks. I hope I've been clear. ;-PP

maurice.makaay on Nov. 3, 2014, 11:25 a.m.
"our uniformity of coding standards and PEP8" Based on that: what's with the lambda? Why not use a def? That would already make this look less scary to beginner Python coders.

veky on Nov. 3, 2014, 2:01 p.m.
OMG. Ok, I understand if my writing is boring to you, but then I don't know what I can write here that I haven't already written in this same discussion. Please read comment <a href="http://www.checkio.org/mission/secret-message/publications/veky/python-3/filter/">http://www.checkio.org/mission/secret-message/publications/veky/python-3/filter/</a>#comment-18873 (from "OK, imagine this scenario"). If there are further questions, feel free to ask.

maurice.makaay on Nov. 3, 2014, 2:07 p.m.
OMG. It does not answer my actual question: why a lambda? Being scary to beginner coders is just a facet that I mentioned in line with the discussion. Feel free to not answer when you find my questioning boring.

veky on Nov. 3, 2014, 2:45 p.m.
I'm sorry about that. What's your question exactly? "Why not use a def"? <ol> <li>Can you answer the question "why not use while loop when you need to multiply elements in a sequence?"? (After all, it would be less scary to some beginners.)</li> <li>And do you realize the answer is basically the same?</li> </ol> Or is your point that we should always use the most general construct?

maurice.makaay on Nov. 3, 2014, 3:15 p.m.
Yes, you can interpret my question in such way that the answer would basically be the same: Pythonic only looks scary when you are not used to Pythonic. For those who know lambda, the code is perfectly readable. But you are misinterpreting my basic question. My basic question still remains: why use lambda instead of a def?

veky on Nov. 4, 2014, 5:43 a.m.
Because it's the most direct way to write what's needed. No, really. Please try to answer: why use for instead of while when you need to multiply all elements in a sequence?

maurice.makaay on Nov. 4, 2014, 6:28 a.m.
Hardly a comparison, but nice try at making it sound like Some Universal Truth. Using * versus mul() is closer to this case I'd say. I'll settle for the answer that you simply see it as the only proper way. I got from another thread that you're seeing lambda as a functional programming tool in Python. You're entitled to interpreting lambda like that and I can work with that premise when reading your one liners.

veky on Nov. 4, 2014, 8:06 a.m.
No, not "only" proper way. But simply the proper way. Ok, have your own example, but mul is simply wrong (it's just a shorthand from the library, <a href="https://hg.python.org/cpython/file/default/Lib/operator.py">defined</a> as *.). If you really wanted something from the language, you might use int.__mul__ or float.__mul__ or complex.__mul__, but here I think it's obvious that most of the time you really need *, if for nothing else, then because your algorithm is invariant to the type of left multiplicand. May I suggest another example, that's been added to Python after it became popular (so you cannot say that it just stayed because of tradition), and it has no real equivalent in other mainstream languages (so it is even weaker than lambda - C# recently got lambdas, IIRC): with statement. Whatever you can do with it, you can do with try and finally, and most of the time in a fairly obvious way. But still Guido went out of his way to make a (fairly complicated) protocol, just for the reason that in most cases where you use finally, you don't need full generality of it, and it helps the reader immensely if it can immediately be seen that "oh yes, this is one of these ordinary cases". So yes, introducing new constructs limiting ones you already have is completely acceptable in Python, when it helps achieve clarity of expression. Even if it is not picked from some other language (and so not intuitively understandable to transitioners), but invented in Python community.

maurice.makaay on Nov. 4, 2014, 6:57 a.m.
Here's a quote from Guido about lambda: "Lambda was really only intended to be a syntactic tool for defining anonymous functions. However, the choice of terminology had many unintended consequences. For instance, users familiar with functional languages expected the semantics of lambda to match that of other languages." That is how I see it too. It's basically a way to define a function, just like def is. The main differences are the anonymity and the limited syntax (although you can do anything in a lambda with some trickery). By binding the lambda to a variable, all that is left is the limited syntax.

veky on Nov. 4, 2014, 7:44 a.m.
Original intent is one thing, language evolution is another. For another example, Guido was at first vehemently opposed to decorators, and now he admits they are incredibly useful. BDFL can change his mind. Believe me, if he wanted lambda gone, it would be gone in Py3 - just as reduce is gone from the core. BTW you're still trying to embed my thinking into one fixed philosophy, in this case functional programming. I assure you it's not how my mind works. I have no problem whatsoever with not being able to have statements inside lambda, just as I have no problem whatsoever with not being able to temporarily go backwards in a for loop. Limitations are useful, since they help reader see what's going on on a generalized level, without tracing each and every token to see if someone hasn't "smuggled" something. Lambda is (delayed) expression with holes. for loop is the simplest iterating device. If you want complicated index management, write it explicitly in a while loop. If you want suites of statements, write them explicitly in a def. The only problem someone might have with it, is "there should be one obvious way to do it" - if someone is a language reductionist, they must feel that lambda is somehow redundant with def. But for is redundant with while, too. And with is redundant with try. And assert is redundant with raise. That's not a reason to banish them. It's a reason to think more carefully what exactly do you need.

Peter.White on March 26, 2015, 7:58 p.m.
<blockquote> Lambda is (delayed) expression with holes. </blockquote> Ahh, now I get it. Now I have a mental image I can work with. I too was somewhat scared of lambda at first, but seeing it being used in so many missions I solved so far, made me wonder why that is. Maybe a little late for a reply, but I just stared yesterday. ;) Anyway, thanks for this and many other educating comments. I really enjoy reading them.

veky on March 27, 2015, 1:40 a.m.
Thank you. I've started wondering if I am crazy, it seemed like almost noone gets my view of the language anymore - that I've converted those that I could a long time ago, and nobody new in a long time. You've restored my faith. :-)

mu_py on Feb. 18, 2023, 7:56 p.m.
<blockquote> and nobody new in a long time </blockquote> well, well - no good deed stays unnoticed ;-) Though I've to commit that it feels quite crazy to comment on those very old discussions....

veky on Feb. 26, 2023, 7:36 a.m.
No big deal, I read all the comments. ;-)

asharkey3 on Sept. 27, 2014, 3:19 p.m.
I put simpler in quotes because I do understand that your solution is a simple one. A single line. That's pretty simple. But a beginner coder is never going to have heard of those commands. So there are definitely easier ways to solve it. It seems I've touched a nerve. Honestly, if being rude and condescending to people is your "persona"..... You need help. You made a ton of assumption of what I am thinking which is just laughable. You are the expert though you would know all of my inner thoughts wouldn't you? Look. I know you're intelligent. You're skilled with the language. That's great. Just don't be a condescending wad of garbage. Am I being CLEAR? Lol.

veky on Sept. 27, 2014, 6:42 p.m.
It seems you haven't read what I've written. Just try to read again the parable of while vs. for. If it doesn't help, sorry.

dalluca on Feb. 29, 2016, 12:27 p.m.
Hi. I am quite new on this platform. I think you are right. And I appreciate your code. I learned how lambda functions, the function filter() and str.join() work tring to understund your code. I think clearity is a relative concept. Basically, the deeper is the Python knowledge you use to build your solution, the more are the persons who do not see your solution as a clear one. I think if you want to write a clear code you should: 1) try to build a as efficient as it is possible function using as basic as it is possible tools. 2) try to build a code that express your solution step by step. Let's try to “code as we speak”. The way to write the same solution in a shorter and more cryptic way is a successive step and maybe it belongs to creative section. On the other side I don't know if it is clever to don't use tools that you know exist. Then, let's do this way: public your solution on both … clear and creative section … everyone will be ok!! :-) I also think that someone who come from another language should know that before to write a Python code he should read the basic concepts of Python language. None can think to program in one language with notions belonging to another language. I hope I was not annoying. I just had some time to spend. Greetings Luca

veky on Feb. 29, 2016, 6:03 p.m.
<blockquote> I learned how lambda functions, the function filter() and str.join() work tring to understund your code. </blockquote> And that's the most important thing. Where you're slightly wrong is in those two points. First one, I know you think it's obvious which tools are basic, but it really isn't. It depends immensely on your previous exposure to basic paradigms. From your code, I see you use for loops and if statements. Both of them are quite advanced concepts for someone with functional programming background: they use recursion and conditional expressions to accomplish similar things. But if it's hard for you to imagine, see this: in Monkey Typing, you use `word.lower() in text`. Try to answer truthfully: why didn't you lower every character separately, and why didn't you code some string searching yourself? You'll probably say "because it's ridiculous... this way is much easier". That's exactly how I feel when someone says `filter` is unclear. It says exactly what it's doing; what else would you want? :-) Second point is much closer to the truth, but irrelevant. I am coding as I speak. In fact, many people tell me I speak Python. :-D Joking aside, language we use does affect how we think, and consequently how we speak. To me it's perfectly natural to say that secret message is a string consisting of all uppercase letters from message. And that's what I wrote. About your further suggestion, you'll surely see that many times I do post a solution in both sections. I do write creative solutions, it's just that you and I don't consider the same solutions creative. For me, "creative" means using Python's features in a way they are not meant to be used. For a cute example, <pre class='brush: python'>defaultdict(print) </pre> is an extremely creative way to get a dict having None values by default. :-D By that definition, this solution is not creative. `filter` and `''.join` are meant to be used in exactly this way.

dalluca on Feb. 29, 2016, 9:24 p.m.
I understund your point of view!! I will try to increment my knowledges to see as much solutions as I can as clear ones :-D That's a good way to grow our own skills!!

famaral on Sept. 19, 2014, 2:31 p.m.
yeap, my bad

jakobEatsCake on Jan. 15, 2015, 8:26 p.m.
this one is categorized correctly it is very clear. Very quick too. Wow one line. I have 7. Impressive. I need to get comfortable with lambda

veky on Jan. 15, 2015, 9:54 p.m.
Thank you. This just shows how people are different. :-) Anyway, about comfortability: whenever you see a <pre class='brush: python'>lambda spam: eggs </pre> replace it with `tempfuncname`, and somewhere above it in the same scope write <pre class='brush: python'>def tempfuncname(spam): return eggs </pre> Whatever spam and eggs might be. So, `lambda` is just a special case of `def`. Same as `for` is special case of `while`. That doesn't mean you should discard it. You just have to be more precise about what exactly you need.

mikaeldovbnia on Feb. 16, 2023, 8:56 p.m.
I agree with @veky. Lambda, filter, and ''.join are very easy and regularly used python functions. This is also the shortest solution I saw. I asked ChatGPT about how to shorten this code, but it said this is the shortest version. This solution is 100% clear.

JosephBernard on Nov. 3, 2014, 8:48 a.m.
I would also call this creative and not simple. Although the below text you posted I would call simple as coming from c/csharp/c++/java I can pretty well logically guess what everything is doing. So coming from any of these I assume from previous programming language knowledge that def just means you are defining something, so you are defining that you want to find a message in some given text. <pre class='brush: python'>def find_message(text): return ''.join(filter(str.isupper, text)) </pre> Also lambda is just a Greek letter with no specific meaning, so unless you look up what it means in python you have no clue as to what it does.

veky on Nov. 3, 2014, 10:51 a.m.
Precisely. But the conclusion is ridiculous. So I should write C++ in Python just so you could understand it? Sorry, it doesn't go. When you come to another country, you learn the language. Of course, if natives are nice, they'll try to speak with you in a simple language, maybe using terms of your mother tongue if they know them, so you could feel more comfortable. But they are expecting you to learn their language, with time. One very important point: c family of languages is not the only one. There are many languages where defining functions doesn't make sense at all. There are more other languages where defining functions is precisely the same as naming lambdas. There are pattern-based languages, where you might think you're defining a function with a piece of code, but it's a wrong metaphor that will bite you sooner or later. (Yes, in Python, a global def is usually similar to C function, but even that doesn't have to be. It might define a generator, for example, which really has no real equivalent to C code.) And saying that "lambda is just a Greek letter with no specific meaning" is incredibly ignorant for a programmer, I must say. Even Wikipedia, which is an encyclopedia for general public, has programming context immediately after linguistic one for lambda: <a href="http://en.wikipedia.org/wiki/Lambda">http://en.wikipedia.org/wiki/Lambda</a>#Lambdaasaprogrammingconstruct

restingsound on Nov. 27, 2014, 5:58 a.m.
I'll jump in on the Creative over Clear debate. I think this solution is mislabeled because it takes the basic form of the checkio puzzle and changes it. The puzzle invites you to change the code within the def, by removing the def you are going outside the normal parameters of the puzzle, which is certainly clever and creative, but not necessarily clear. The fact that it reads clearly to you or some people isn't really what needs to be debated. I think you should be proud of your Creative solution, don't sell yourself short.

veky on Nov. 27, 2014, 8:58 a.m.
I should be proud, right? And that's the reason you've sprinkled -11 thumbs over my various comments in the span of few hours? Or is this completely unrelated to that? Your thumbs, your life... but I want to tell you you'll be taken seriously much more often if you weren't so bitter. <hr /> And now let's try to address your comment. I'm really perplexed: where does the puzzle invite me to change code within the def?? I must say I don't see it anywhere. The task says: <blockquote> You are given a chunk of text. Gather all capital letters in one word in the order that they appear in the text. </blockquote> Yes, it says a lot more, but I must say I don't see what you wrote. Why did you get such an impression??

restingsound on Dec. 2, 2014, 5:32 a.m.
I got the impression from the first problem I looked at, and the same sort of comments written for other problems (although not explicitly on this one): > <blockquote> def checkio(number): #Your code here #It's main function. Don't remove this function #It's using for auto-testing and must return a result for check. <pre class='brush: python'>#replace this for solution return str(number) </pre> </blockquote> I've given you -thumbs when I don't like what you have done whether it be categorization, style, content, or attitude. I believe that is the point of the button being on the page. I think its a much more positive method of debate than name calling and passive aggressive emoji, but to each his own.

veky on Dec. 2, 2014, 10:06 a.m.
<blockquote> I got the impression from the first problem I looked at, and the same sort of comments written for other problems (although not explicitly on this one): </blockquote> Well, on the starting tasks, people probably need someone to hold their hand. As tasks progress, you get more and more freedom. If you gave me -1 with that reason on that task (I think it's from Fizz Buzz, right?), it would make a bit more sense. But still not much, since I don't view the default code as part of the task text. It is there as a "template hint", to show you how your code is supposed to look. Besides, default code changes a lot... how do you know it was there when I solved the task? :-) And "removing a function" is debatable... I have a function callable as checkio, it's just not aware of its name. But ok, I accept your explanation, though as I say, it would help if that -1 was on the task that explicitly says not to remove the function. Now on to a more interesting thing: <blockquote> I’ve given you -thumbs when I don’t like what you have done whether it be categorization, style, content, or attitude. I believe that is the point of the button being on the page. I think its a much more positive method of debate than name calling and passive aggressive emoji, but to each his own. </blockquote> Well, "positive" doesn't mean much (especially when votes are negative:)). It surely isn't a constructive method of debate. Believe it or not, I care what people think about my style of teaching. I am aware a lot of people don't like it, but also a lot of people do like it very much. To tweak it precisely, I must know what bothers those who don't like my lessons. (You're not obligated to give me that information, of course, but then don't expect much understanding in return. Also, I'm not obligated to change my behaviour because of that info, but then you're entitled to continue to think of me as a jerk.:) Those who like me, usually tell me the details. (Some of them who don't like me also do, and I'm thankful for that.) But a rain of minuses in one hour doesn't really convey a consistent message, especially if, as you say now, those minuses are for wildly different reasons. <a href="http://www.checkio.org/mission/non-unique-elements/publications/veky/python-3/shortest/#comment-18876">Here</a> is an example, a relatively recent one. I honestly ask what's the problem, and I get -1. And no reply. Ok, give me -1 if you want, but still answer my question. Is it so hard? Now let's try to address your reasons. <ul> <li>Categorization: you probably know my reasons why I think categories are mostly missing the point as currently implemented, and I mock them. But that's addressed to the admins. Towards the users, I always try to explain my thinking, and many times I've been convinced to change the category.</li> </ul> In this concrete mission, there is a mismatch between two meanings of "solution": process of solving, or its result. Though I agree my solution in first sense is more creative than clear, I firmly hold the opinion that solution in second sense is more clear than creative. Up until now, I always thought "solution" on CiO meant the second definition. (You opened my eyes to the first possibility, and thank you for that.:) <ul> <li>Style: might mean discussion style or coding style. First one I think can be put under "attitude", so it is addressed there. Here I'll try to talk about coding style. As you probably know:-), I value Spartan programming, but more importantly, I value capturing right abstraction levels.</li> </ul> If something can be done in two ways, almost always in Python one of those ways is a generalization of the other (for vs while, with vs finally, enumerate vs range, lambda vs def, import vs __import__, module vs class, Enum vs SimpleNamespace...). In such situations, e.g. Java values always writing the most general thing. I think, and have good arguments for it, that Python values the opposite approach: if x is A, but it is also a B for a trivial reason that all As are Bs, then if someone asks you what x is, it is much more useful to say it is A than it is to say that it is B. Python is about explaining algorithms (putting them to mind, so they can be understood), Java is about implementing them (putting them to work, so they can be maintained). Of course I'm oversimplifying, but the general idea is in that direction. I think, but of course I can't be sure since -1 alone doesn't speak much, that all of your dissatisfaction with my coding style stems from the fact that I adhere to the above principle (sometimes fanatically, yes:). I'm not willing to give up on that principle, but if there is some other dissatisfaction with my coding style, I'd like to know. <ul> <li>Content: well, this hurts me most. In my opinion, I try extremely hard to write things that are explanatory and at the right level of detail, yet still aren't technically wrong. Python's consistency makes it possible at all, but it is still hard to do. I'd really like to know the details here: what substance have I written here, that you didn't like?</li> <li>Attitude: strangely, that doesn't hurt me at all. :-D I'll try to explain. I've been in many discussion forums, and here is a principle I've noticed almost always works: if the forum is about foxtrot dancing, then everything about foxtrot is on topic. Everything about dance in general is also welcomed, probably. Discussions somehow related to dancing are ok, though might be redirected somewhere else. Discussions remotely related to dancing might theoretically be tolerated. But discussions about participants: about someone's character, attitude, demeanor, use of words, and everything else in that area; are strongly not ok.</li> </ul> That surely doesn't mean "anything goes". If you have a problem with someone's attitude, you can (you're even encouraged to) report them to moderators, who will decide what atmosphere they would like in a discussion. Or if the problem is subjective and affect only you, most forums enable some kind of blocking. But that's not the topic of discussion. If forum is so huge that such things can be meaningfully discussed (that is, moderators are so numerous they don't always agree on an atmosphere), there is a special separate meta-forum for such discussions. Conflating these things is almost always a very bad idea, for many reasons, some of which have been scientifically proven. In a few occasions, I have been contacted by moderators about my style of talking, and I tried to change it to be more "in line". Since that was a long time ago, I assume I have been successful. :-) But that things have to be discussed privately, or at least separately from the main topic. You might argue that your -1s are precisely that: a separate channel on which you can signal your dissatisfaction with me as a person. But I doubt other users view it that way (from what I've seen, even you conflate giving minuses for attitude and for coding style), and more importantly, I doubt moderators view them that way. You'll probably be much more successful if you contact them directly.

Rat_Eater on Jan. 23, 2023, 4:57 p.m.
<blockquote> I must know what bothers those who don’t like my lessons. </blockquote> Just by reading a few of people's comments and your reactions, i've found multiple occasions where people explain very clearly what bothers them, but instead of you showing understanding in return, <blockquote> You’re not obligated to give me that information, of course, but then don’t expect much understanding in return. </blockquote> you immediately start defending yourself, bringing up your lvl, your 'high' follower count, calling them names. (meanwhile in another comment, you asked somebody why they like followers so much "This is not instagram") <blockquote> but then you’re entitled to continue to think of me as a jerk </blockquote> hmm wonder why that is. When people express their dissagreement, you have a couple options, you can either: <ol> <li>Ignore (very easy, costs no time or energy)</li> <li>Respecting their opinion and calling it 2 solutions to the same problem (rather easy, costs little time)</li> <li>Acknowledge their solution but proving, with strong supporting arguments, that yours is superior (little difficult, time consuming)</li> </ol> Instead you seem to love option 4 <ol> <li>Insist on your solution being best without giving solid reason (very easy, costs little time, makes other people see you as a self-proclaimed genious)</li> </ol> That exactly is why you receive so much -thumbs, negative comments, etc. people don't like I don't know how you are now (after 8 years), but i hope, for your sake, that you're not still like this. If you are, then i am very sorry for all those who live in your proximity. Thank you :-)

veky on Jan. 23, 2023, 8:35 p.m.
<blockquote> Just by reading a few of people’s comments and your reactions, i’ve found multiple occasions where people explain very clearly what bothers them, </blockquote> ... yet for some mysterious reason you can't give even one example here? :-) [very easy, costs no time or energy, right?] <blockquote> but instead of you showing understanding in return, </blockquote> Yeah, showing empathy is where I'm weak. I know that. But you can be sure that I understand most of them. <blockquote> you immediately start defending yourself, bringing up your lvl, your ‘high’ follower count, calling them names. </blockquote> Since you haven't given examples, all I can give is general view. And that is, roughly, don't believe that your view is the only one that counts. For every person that hates my solutions, there are at least three that love them and learn from them. And it's much healthier and more useful to focus on providing value for those people. If I can fix something concrete about my solutions that makes that person who hates them, hate them less, I usually do that. (You can probably see the evolution of my style here through 10 years.) But if it compromises what I'm trying to accomplish, then no. It isn't worth it. <blockquote> (meanwhile in another comment, you asked somebody why they like followers so much “This is not instagram”) </blockquote> A subtle difference. I think I remember that situation. That person was doing concrete actions that don't make much sense in this environment, only to get more followers. I told them that being motivated by follower count is probably a bad idea (as opposed to instagram, where that's the main motivator by design). If you do actions that make sense to you, you would do them even if there were no follower count, and as a consequence you get a lot of followers, then that usually means you're doing something right. <blockquote> hmm wonder why that is. </blockquote> A few explanations come to mind. :-) <blockquote> When people express their dissagreement, you have a couple options, you can either: Ignore (very easy, costs no time or energy) </blockquote> Actually not for me. But it's a complicated topic and not relevant here, since I don't think I ignored any such comment anyway. <blockquote> Respecting their opinion and calling it 2 solutions to the same problem (rather easy, costs little time) </blockquote> I'm not quite sure what this means. For many missions here I do have 2 solutions, or even more of them. Precisely because people learn in different ways, and not everybody must like a particular style. Even when I'm not the best person to write a solution in a particular style, somebody else is. I never forbade any of my followers to like somebody else's solution. Let the pluses speak for themselves. <blockquote> Acknowledge their solution but proving, with strong supporting arguments, that yours is superior (little difficult, time consuming) </blockquote> I'd cross out "strong" (since usually it's not of much importance that I invest a lot more time in strong arguments, people who I can convince will usually be convinced with any arguments, and people who I can't convince won't usually be convinced with any arguments either). And more importantly, I'd add "in a particular context" after "superior". Of course every valid solution has some advantages over others. (Even invalid solutions might have advantages. The world is complicated.) If I point out that solution A has better x and y, while solution B has better z, that just means that if you value x and y more, A is probably superior. If you value z more, then it isn't. <blockquote> Insist on your solution being best without giving solid reason (very easy, costs little time, makes other people see you as a self-proclaimed genious) </blockquote> ... again, no examples? :-) I wonder why. <blockquote> That exactly is why you receive so much -thumbs, negative comments, etc. people don’t like </blockquote> Ratio of negative to positive comments/thumbs I receive is about 1 to 10. "so much" is always relative. :-) <blockquote> I don’t know how you are now (after 8 years), but i hope, for your sake, that you’re not still like this. </blockquote> I'm sorry, but on CiO I'm pretty much the same. As much as you think it hurts others, I really don't see what you mean by "for my sake". Trust me, I feel fantastic here. I wouldn't be here for almost a decade if I didn't, of course. <blockquote> If you are, then i am very sorry for all those who live in your proximity. </blockquote> What makes you think that those who live in my proximity have any idea of my life here? You react the same on social networks and in real life? :-) <blockquote> Thank you :-) </blockquote> For what?

dig on Jan. 23, 2023, 10:53 p.m.
Please Veky, could you follow me in order to see all your codes? I could learn a lot:)

veky on Jan. 24, 2023, 3:26 a.m.
I'm not sure I understand. Follow you where? Whose codes? I think it doesn't work that way. If you follow me, then you can see my solutions as soon as they are published. If I follow you, then I'll be notified about your solutions when you publish them. But I can see your published solutions anyway, just by going to your profile page. I just did that, and there I saw something interesting (as I usually do). Please re-read <a href="https://py.checkio.org/mission/secret-message/publications/veky/python-3/filter/share/efd4e41ef30dcac9af99dcd6c84ae149/#comment-18873">this comment</a> starting with "Ok, imagine this scenario". Let me requote the relevant part. <blockquote> you have to multiply all the elements in a sequence. Of course, the clearest way to do it is [a for loop]. But is it obvious? Apparently, from what I see during teaching, and here on checkio, it isn’t (for everybody). If someone comes from some unnamed programming languages, while loop is more obvious. Now imagine he says: "this is creative because there are far simpler answers for a coder like me. I myself don’t know what for loop does". And he’s right. He’s not trolling (this is important!). Of course, he does know what for loop does in C (oops, I named it:), but that doesn’t really help you understand what it does in Python, unless you think on a very high level of "iterating through a container", that many programming beginners don’t. For him, Python for loop is something mysterious, and he sticks to what he knows, because while loop, although wordy and convoluted, at least works the same way (at a first glance at least) it works in Java (oops, named it again;). </blockquote> Recognize yourself? Look at <a href="https://py.checkio.org/mission/adjacent-letters/publications/dig/python-3/first/">this one</a>. :-) Rarely I see such a direct example for my claims. Thank you. :-D

dig on Jan. 26, 2023, 9:10 p.m.
Perfect, now i understand how it works (following part). About the other thing, I guess you wanted to reply another user. Thanks Veky, I will be expecting about your solutions:)

veky on Jan. 27, 2023, 5:11 p.m.
<blockquote> I guess you wanted to reply another user. </blockquote> Not really, I think you can learn from it too. :-)

melizzaryan on May 25, 2015, 12:35 a.m.
veky i think you're the bitter one... so negative and condescending especially for being on the beginners section!

veky on May 25, 2015, 4:43 a.m.
Please read the "Attitude:" section in the <a href="http://www.checkio.org/mission/secret-message/publications/veky/python-3/filter/share/efd4e41ef30dcac9af99dcd6c84ae149/#comment-21350">comment above</a>. (You might also read the whole comment, if you want to understand my motives better. But that's optional.:)

CounterSparta on Nov. 29, 2014, 10:02 a.m.
No need to use 'join' here, 'filter' returns list, tuple, or string depending on the given sequence type. So this will work with less code: <pre class='brush: python'>filter(str.isupper, text) </pre>

veky on Nov. 29, 2014, 3:13 p.m.
In lesser pythons, yes. But in True Python, no. :-) Anyway, nice find. BTW you don't have to define a new function at all, partial will do it for you. :-) <pre class='brush: python'>find_message = partial(filter, str.isupper) </pre>

CounterSparta on Nov. 29, 2014, 5:47 p.m.
Thank's for suggestion and 'partial' tip, didn't knew <pre class='brush: python'>filter() </pre> behaves in other way in python3.

veky on Nov. 29, 2014, 8:39 p.m.
The main idea was to enable "true piping" of data, not concretizing every intermediate result in memory unnecessarily. That, combined with desire for explicitness, resulted in having to concretize the result back into value-land manually ("".join to produce a str). Inconvenience for simple scripts, but really much better and faster in big systems.

marshall.zheng on Dec. 24, 2014, 8:08 a.m.
studied lambda

veky on Dec. 24, 2014, 10:23 a.m.
Finally! :-D

Hayertjez on Nov. 14, 2015, 11:09 a.m.
What is the resource you used studying. At this moment I find this function unclear.

veky on Nov. 14, 2015, 1:30 p.m.
What exactly? Lambda? Does <a href="http://www.checkio.org/mission/secret-message/publications/veky/python-3/filter/share/efd4e41ef30dcac9af99dcd6c84ae149/#comment-33218">this comment</a> (third paragraph) help?

Hayertjez on Nov. 16, 2015, 5:27 p.m.
Bookmarked, thanks!

SrdjanTubin on Dec. 30, 2014, 10:55 a.m.
I would call this solution creative, then clear. When we have a code like this, as a part of a much bigger software component, code readability is reduced, which is...not really good (because, in software maintenance for example, code readability is very important!).

veky on Dec. 30, 2014, 6:25 p.m.
It depends on what do you exactly need to maintain. May I ask you a question? You need to add all the numbers in seq. Would you write <pre class='brush: python'>total = 0 for number in seq: total += number </pre> or just <pre class='brush: python'>total = sum(seq) </pre> ? And what of those two do you find more readable? Then we can continue discussion.

SrdjanTubin on Dec. 31, 2014, 8:03 a.m.
Anything you need to maintain or extend after one or more years, is as easy to do as it is readable :) (ofcourse, many more parameters play a role here, but code readability is one of the most important). "total = sum(seq)"is FAR more readable, i agree with that! But, you could make your solution (in the original problem) a bit less creative, break it in 2 or 3 lines of code, but more readable.

veky on Dec. 31, 2014, 9:10 a.m.
You really don't see you're a bit contradictory here? When I ask whether it would help readability to break sum in 2 or 3 lines of code, you agree that it would be pointless. Ok, let's make one more step. You need to add all strings in seq. Would you write <pre class='brush: python'>ans = '' for string in seq: ans += string </pre> or <pre class='brush: python'>ans = "".join(seq) </pre> ? (You can't not see where I'm going with this. But if you're willing to continue to pretend, that's fine with me.:) <hr /> If you don't want to play stupid A/B games, I understand. But I'm trying to prove to you that this solution is not creative at all. I'm using builtins in exactly the way they are meant to be used, and capturing algorithms on their natural abstraction level, backed by Python concepts.

yaab22 on Jan. 5, 2015, 11:27 a.m.
Can you explain a bit about the use of the unbound "str.isupper"? I understand the meaning of course, I just wouldn't know how to use this unbound syntax myself... I guess it is like calling the class internal method, but without any specific object?

veky on Jan. 5, 2015, 2:21 p.m.
It is called with an object, as opposed to ordinary method calls, which are called on an object. ;-) For "normal" instance methods, <pre class='brush: python'>x.meth(*args) <~~~~> type(x).meth(x, *args) </pre> For example, if you have `s=[4,5,2]`, then instead of `s.append(3)`, you can write `list.append(s,3)`. And instead of writing `"blah".isupper()`, you can write `str.isupper("blah")`. Clearer?

yaab22 on Jan. 6, 2015, 1:54 p.m.
Clear - many thanks. Is there some guidelines to prefer this syntax over others? I'm not looking for a black or white answer, just interested in the reasoning for this wide variety of possible syntax. I'll be glad if you can refer me to some reading, if you prefer. And thanks again for your explanations!

veky on Jan. 6, 2015, 7:59 p.m.
<blockquote> Is there some guidelines to prefer this syntax over others? </blockquote> It depends on what the others are. :-) filter expects a predicate (ordinary function of one argument), and methods without parameters are (almost) the same as them. You can use pointful syntax with comprehensions, but it is not precisely the same generally (filter evaluates its arguments at call time, yet applies the predicate at iteration time). And in the case of filter, it is even more cumbersome since you have to write the point name three times (as opposed to usual two). <pre class='brush: python'>"".join(point for point in text if point.isupper()) </pre> But I think you meant something else by "others". What? :-) If you just want to read something about how it all came to be, and you have a lot of time, I suggest <a href="http://python-history.blogspot.com/.">http://python-history.blogspot.com/.</a> From the horse's mouth. :-D (Just be careful, Python has evolved much since those were written. But if you want to know the traditional aspect, there it is.)

DavidJones on Jan. 6, 2015, 2:27 a.m.
If you're going to viciously defend yourself for being clear, can we get a new tag that implies a more "beginner" approach?

veky on Jan. 6, 2015, 7:51 a.m.
No problem with me (however, I'm not the right person to ask that). But "beginner" doesn't necessarily mean what you think it means. Java way of solving such things (what you probably mean, since Java-obsessed people usually hate solutions like these) is just one of many approaches. Beginners to Python come from many different backgrounds, and although C-family emigrants are in the majority, the other families of languages are not negligible. Here are a few approaches: <pre class='brush: python'>find_message = functools.partial(filter, str.upper) </pre> (mentioned above, Python 2 style) <pre class='brush: python'>def find_message(text): b = bytearray(text.encode()) i = j = 0 while j < len(b): if 64 < b[j] < 91: b[i] = b[j] i += 1 j += 1 del b[i:] return b.decode() </pre> (ANSI C, K&R would be proud) <pre class='brush: python'>def find_message(text): try : x, *xs = text except: return "" ys = "".join(find_message(xs)) if x.isupper(): return x + ys elif True : return ys </pre> (Haskell newbie) <pre class='brush: python'>def find_message(TEXT, ERASE=int()): if ERASE.__eq__(256): return TEXT if chr(ERASE).__eq__("A"): ERASE = ord("Z").__add__(1) return find_message(TEXT.replace(chr(ERASE), str()), ERASE.__add__(1)) </pre> (COBOL renegade) <pre class='brush: python'>from functools import reduce as foldr find_message = lambda text, remove_if_not=filter: foldr( str.__add__, remove_if_not( list(map(chr, range(*map(ord, "A[")) )).__contains__, text ), str() ) </pre> (LISP hacker) Which one do you think is "beginner" approach? :-) I probably exaggerated a bit since the task is so simple, but the point stands: Python beginners are extremely influenced by their previous experience, and are always trying to speak their native language in Python (and Python lets them, usually:). And most importantly, Java is not the only possible entry path. ;-)

DavidJones on Jan. 7, 2015, 2:35 a.m.
How about that I don't have any background? What snark would go best with my level of ignorance. Is this website simply not aimed at anyone without previous coding experience? I'm trying to learn with codecademy and this, but you are seriously the worst community member I have ever met. I hate coming in here because of you. Python was supposed to be easy and fun but instead it's become a battle to try to learn. So if this website was aimed to be all inclusive, you've successfully managed to stop that from happening. Most people here with questions (like mine) seem to be coming from my background (i.e.: none) and you've never once offered to explain how python works. You just throw annoying and snarky code at us. So yea, you're right. Good job [the guy i don't like] =moderated=.

veky on Jan. 7, 2015, 9:31 a.m.
I appreciate your comment, though probably it will be moderated because of some inappropriate words. So you're saying Python is the first programming language you ever learned? That's really hard for me to believe, but let's suppose so. There still has to be some way you think about programs. Maybe as pseudocode (a.k.a "plain English", in elementary steps). Python allows you to write that, too. <pre class='brush: python'>def uppercases(text): # to collect uppercases from text for letter in text: # go through all letters in text if letter.isupper(): # and if a letter is uppercase, yield letter # collect it def find_message(text): # to find message in text, collected = uppercases(text) # first collect uppercases from it, result = "".join(collected) # join them with nothing in between return result # and return the result of that </pre> [About my attitude: sorry, I won't discuss that. I will just urge the admins to please make some kind of a blocking feature, so you can continue to come here and not have to read my posts. I'm sorry you feel that way about me, but I'm not willing/able to change.] However, I must react to this piece: <blockquote> you’ve never once offered to explain how python works. You just throw annoying and snarky code at us. </blockquote> First part is patently false (I've written a huge body of comments here about internal workings of Python, I'm sure you'll find them if you look around), and the second part is just a misunderstanding. You see, I wrote a solution. I insist it is clear to people who know Python. It cannot be clear to people who don't know Python, but no solution can be clear to all the people who don't know Python, simply because they have many preconceived notions about how code (or algorithms, if they haven't worked with code before) should look. That's why CheckiO is so great, after all: you don't see just one approach, you see enormous number of them. I'm not the only one here. And of course if people saw only my solution, I would write it very differently. But still I couldn't write it for everyone. So this system is surely better. And then you come and ask "can we get a new tag that implies a more “beginner” approach?". I (seems incorrectly) assumed that by "tag" you mean a new category for solutions. I tried to explain to you that such category, even if it existed, would not help you in a way you think it would. Beginners don't learn best from looking at other beginners, especially in coding. I read beginner code all the time, and the only common characteristic I find is that they are incredibly diverse and use approaches no sane Pythonista would use. I'm aware expert code looks alien to you, but I'm convinced many other beginners' code you'd find even more alien. If I read it correctly now, it seems you wanted a real tag, not category, but not for solutions, rather for some new part of CheckiO, where someone would explain, in a lesson style, how one should go about finding a solution. Currently such a part doesn't exist. CheckiO has "hints" for that purpose, but if I read you correctly, you find them insufficient. If CheckiO opened some new part of site in that style, I'd probably write there. But I'm happier with current configuration, where I can solve, publish and review. And I suppose many users here are also, simply because that's what makes CheckiO different from other communities of that type. You mentioned Codecademy... I can't stand it and would probably never be really active there, but it can be helpful to you (considering how bad a person you think I am, probably it's better for you that I don't go there often:)). Various communities have different goals, and you're free to choose what suits you best. <hr /> The TL;DR point is, you haven't asked how a specific piece of code works (if you did, and I saw that, I'd probably invest much time and effort to explain, as many people here can witness). You asked, roughly, to make more visible how a beginner would solve this task. Those things don't have much in common. :-/ If you're willing to give me a chance, try it. Find a piece of code in some solution, and put a link here. I'll try my best to explain. I can't guarantee you'll like it, but I can guarantee you'll learn something new. But you have to ask what you want, not expect others to guess what you meant.

OMFGNuts on Jan. 9, 2015, 10:02 a.m.
Unfortunately, blocking feature won't come out any time soon, we have something more important and interesting coming. The best option for @DavidJones would be to close his eyes when he sees your comments. Spread Python-love and Python-happiness!

JjHack on Feb. 12, 2015, 4:03 a.m.
Relative beginner here. Just for kicks, I tried simply <pre class='brush: python'>filter(str.isupper, text) </pre> and got an error message ("str" object required, got unicode). Can you briefly walk us through what the difference is? Thx.

veky on Feb. 12, 2015, 5:46 a.m.
Well, the real solution is to use real Python (not a milk snake named Python 2). But in case you have to use the milk snake, obviously CheckiO has switched input here to be unicode. I suppose you can just <pre class='brush: python'>functools.partial(filter, unicode.isupper) </pre> then.

drobotnik on April 10, 2015, 10:18 a.m.
WOW! Shots fired in the comments section!!! I think clear for most people means something like this: def find_message(text): <pre class='brush: python'>"""Find a secret message""" output = '' for l in text: if l.isupper(): output += l return output </pre> Either way, I am definitely more motivated to learn lambdas now!

veky on April 10, 2015, 1:34 p.m.
I think it's obvious that you value disassembling to some elementary steps, and that's ok (it's not the only valid view, but it's defensible). But yours is just a very specific view of what "elementary" means. Why is `l.isupper()` not written "simpler"? Why is `+=` not written "more explicitly"? Why is the concept of for loop "clear", and specifically, why iteration through text means iterating through 1-character substrings of it? And on another level: How could += work when output is immutable? Does it copy too much? What is algorithmic complexity class? What about Schlemiel the Painter? Does it, o horrors, depend on the implementation? :-P BTW: don't learn about lambdas. Rather learn about filter. ;-)

drobotnik on April 10, 2015, 3:31 p.m.
Cheers, thanks for the advice!

mauricioabreu on April 21, 2015, 2:54 p.m.
Mine is just: <pre class='brush: python'>return filter(lambda x: x.isupper(), text) </pre> No "join" at all. :-)

veky on April 21, 2015, 2:56 p.m.
You can have no lambda at all, too. And no def/return at all, even. See the above comments. ;-)

mauricioabreu on April 21, 2015, 3 p.m.
Yeah, you are right! Actually your solution is for Python 3.x Python 2.7 filter does not return a filter object. Good solution! :-)

veky on April 21, 2015, 3:06 p.m.
Thanks. I write Python 3 preferably, but this is one of few cases where Py2 DWIMming did the right thing, and Py2 deserves credit for that. :-)

Ursidae on May 2, 2015, 4:20 a.m.
I don't feel that using lambdas in the Elementary section is "clear".

veky on May 2, 2015, 4:42 a.m.
What exactly is the problem here? You're afraid beginners will learn something? Numerous times, people have told me they didn't really get lambdas until they saw my solutions. And the reason is completely lame: not because they wouldn't understand them, but because noone showed lambdas to them, thinking they wouldn't understand. Of course my solution is not "official", it is just one of many. If you feel it's impenetrable, just look at another one. But it's not the problem here, right? You do understand what lambda does, you just have a feeling someone else wouldn't. BTW it's funny you are completely ok with using filter ["comprehensions are better!"], unbound str. method ["you have to understand Python's object model!"], inlining ["it doesn't read smoothly!"], and ''.join ["it being called on a str instead of a sequence is just WRONG!"], all of which have at one point been proclaimed as complete mystery to beginners, usually with comments in brackets as justifications. But no, they are all ok to you - except lambda. Nonsense.

Luke123 on May 27, 2015, 7:39 a.m.
I like the solution, this is the best part of Checkio, to learn think in new ways. First i solved the questions with for loops, like most would do in more low level language like C or C++. Then i saw that most of the solutions was on one line but i didn't understand them, so i checked it out and learned about list comprehension. Now i'am seeing your solution, its mind blowing! Thanks for the sharing!

veky on May 27, 2015, 7:08 p.m.
<blockquote> this is the best part of Checkio, to learn think in new ways. </blockquote> Exactly. :-) It would be incredibly boring if we all wrote the same solution.

LokiOfFire on June 20, 2015, 4:57 a.m.
This is a leaning site. Not a "I'm better than you site." Your solution was not helpful and neither was your attitude

veky on June 20, 2015, 5:21 a.m.
I'm sorry for that. I said I'm trying to persuade admins to make a blocking feature, but it seems they don't want to do it. Please try to understand that my solutions, and my style of teaching and explanations, have helped an enormous number of people here. You can see it in my rank and the number of likes, comments and followers I have. I know it isn't suitable for everybody, but it is really not the end of the world: you can always look at some other solution if you don't like this one, right? BTW are you sure my solution isn't helpful? Have you tried to understand it? You don't have to, of course, but don't have to write comments like these, too. If you're motivated for one thing, you might be motivated for another. ;-)

Czarinov on July 29, 2015, 10:46 p.m.
I've just started out with the site. Python is kinda new to me, I had some programming in different languages in university, but it's been some time and it never was Python. I'm really surprised people give you shit for posting interesting solutions. There is so many other solutions for users to look into. For me personally your solution (mine used isupper and ForLoop) is just another view and new stuff to learn. So I will dig into how lambda works now and try to decipher the rest of your one liner ;)...

veky on July 30, 2015, 6 a.m.
That's the spirit. :-) BTW, people "giving me * for posting interesting solutions" are a small minority. It just seems bigger since CiO doesn't require comments for positive thumbs, just for negative ones. But you see that I'm level 19, so obviously a great number of people has liked my solutions very much. Yes, I'm always saying: CiO's main strength is diversity. You don't see one official solution, you see tens and potentially hundreds of them. And through voting, some of them are crystallized towards the top. Some are "endorsed". But none is exclusive. My solution also uses .isupper, just as a function (of a str type), and for loop is done using a filter, that just picks objects from a sequence satisfying that function. How lambda works is very simple: whenever you have a function `blah` defined as <pre class='brush: python'>def blah(whatever args): return expression </pre> you can replace `blah` with <pre class='brush: python'>lambda whatever args: expression </pre> Helps when functions are simple and needed only once.

lena-py on Aug. 8, 2015, 8:15 p.m.
Thanks for the exposure to "filter" and unbound methods! ps: are thanks redundant when thumbs-up are given? In my line of work cursory "thanks" emails just take away precious seconds of the day. I"m happy to assume that everyone is automatically thankful of my help and prefer to assume it rather than have to take the time to read it.

veky on Aug. 8, 2015, 8:33 p.m.
<blockquote> ps: are thanks redundant when thumbs-up are given? </blockquote> Pretty much, yes. But in my view it's still better than thanks (or any generic praise) without thumbs, which I receive a lot. :-) [If everyone who wrote how wonderful my solutions/posts are gave me just one thumb, I'd probably already be level 20.:] Don't get me wrong, personalized messages are always appreciated. But "cool solution, man" and "thanks for the info" are much better expressed as thumbs/pluses. I receive about 20 CiO notifications a day (with a big deviation... some days it's even a three-digit number), and it helps if I don't have to clickthrough to a comment just to see a generic "you rock". :-) In fact, your comment is on the fence: it's pretty generic, but still, it's nice to know what exactly you've found most valuable in the solution. The problem is, people as considerate as you will probably stuff at least a bit of information in their messages - it's people who never even think about the other side (reading 100 generic comments in a day), who write totally generic stuff. Oh well, it's not a bad position to be in, after all. :-)

lena-py on Aug. 8, 2015, 9:05 p.m.
Thumbs and votes it is, then; unless I feel the need to be explicit.

fasatrix on Aug. 8, 2015, 9:25 p.m.
This is simply beautiful @veky . I had to go and check how filter works to fully get it, but I am glad that you post solutions like this one as they push me outside of my comfort zone, which is where one learns. @DavidJones I know how you feel with @veky as I felt the exactly the same way, but try to remain focused on "what he writes" rather then "how he writes it".. his knowledge (of python) is high and he is keen to share it, which is great if you think about it.. to remain on topic, the filter(function,iterable)is a very powerful tool.. infact, from the official doc, it is equivalent to [item for item in iterable if function(thing)]... literally this reduces to ash any list comprehension... in this specific case being str.isupper our function, each "item" in our "text" is passed onto the function which checks if there is any Uppercase letter. If there is, it will be "joined"...

veky on Aug. 9, 2015, 6:22 a.m.
<blockquote> infact, from the official doc, it is equivalent to [item for item in iterable if function(thing)]… literally this reduces to ash any list comprehension </blockquote> No, no, no. First and most important, read the right documentation. This is Py3, not Py2. So there is no list comprehension anywhere - what they wrote is a generator expression, which is really a different beast. Concretely, it doesn't store values it iterates through in space, but in time, trading multi-pass and random access for memory saving and potential infiniteness. Second, not every list comprehension (nor every generator expression, of course). Only if you already have or can nicely name a function that will filter your iterable. If your first filter argument is lambda, you're doing it wrong. <pre class='brush: python'>filter(lambda item: condition, seq) ~~~> (item for item in seq if condition) </pre> Third and most subtle: while being "mostly correct", docs are not completely true (don't hold it against them: <a href="http://www.checkio.org/mission/flatten-list/publications/veky/python-3/could-be-shorter/share/b75543f35c946baeced2b5bbad69d730/#comment-12644">as I said</a>, in a fantastically dynamic language, docs are always just an approximation of the code). See this for a weird example: <pre class='brush: python'>def f(x): globals()['f'] = bool return x < 5 </pre> Now, <pre class='brush: python'>>>> print(*(x for x in (0,3,7,0) if f(x))) 0 3 7 </pre> but (redefine f as above again) <pre class='brush: python'>>>> print(*filter(f, (0,3,7,0))) 0 3 0 </pre> The point is: f is looked up and resolved only once when filter is called, as opposed to being a normal part of evaluation of generator expression (resolved every time it is iterated through). Of course, probably noone would write such convoluted code - but more natural examples can easily be found. For example, f can be a bound method, whose `self` may change between construction of a filter and iterating through it. It's important to know that while list comprehensions stuff resolving and iterating together, and genexps delay it all to the point of iteration, filter (and map, and similar functions) are a nice in-between model that in many cases we really want. See <a href="http://www.checkio.org/mission/clock-angle/publications/veky/python-3/there-and-back-again/share/77e46ac8af8d058bdb0fdac01be61796/#comment-20518">this</a> for more details.

shuwz on Aug. 27, 2015, 11:46 p.m.
I don't know why people are attacking you. I enjoy coming to your solutions and learning something new every time. I just started a couple days ago. Tip for all the beginners here: Spend less time attacking veky and arguing with him, instead spend the time to learn about his solutions instead. Maybe after you read a couple documents you would realize why this is labeled clear. Obviously it wont be clear if you don't know the methods he is using.

veky on Aug. 28, 2015, 3:59 a.m.
As I said many times, I wouldn't have the points that I have if a lot of people didn't like my solutions. It's just that CiO requires comments for negative votes, so when someone looks at the comments, the situation is biased towards criticism. Also, an important addition to your second paragraph: no style of writing Python is clear for every person who doesn't already know Python. It's just that people coming from C family are in the majority, so they think that what's clear to them would work for everyone. It wouldn't. :-)

anonymorf on Aug. 31, 2015, 12:43 p.m.
Hi Veky, Could you please review my solution ? I wonder whether is it ok to import a function in this case. I'm a beginner and don't want to get bad habits. Thanks! <pre class='brush: python'>def find_message(text): from re import findall x = findall('[A-Z]+', text) x = "".join(x) return x </pre>

veky on Sept. 1, 2015, 1:03 p.m.
I meant for you to post me the link to your solution here, not the solution itself. But nevermind that for now. Importing a function if you only use a single function from module (or importing many functions if you only use one module) is perfectly fine. It's when you use more different modules, when namespacing really show how "one honking great idea" they are. Moreover, there is a slight switch in style of writing modern Python, to always use `from...import` if possible, since it is a bit more general (for example, it allows relative imports). Then `import string` is just understood as a shortcut for `from stdlib import string`, where `stdlib` is a mythical top-level namespace for builtin modules. But it's possible that this is just a fashion thing, and that it will fade. Of course, it should be said that `from module import *` is usually frowned upon (except in some old modules which do their own namespacing), and anyway, couldn't be used here legally since from Py3K there is no real way to have dynamic local names inside functions. [There is another thing just to be complete. Guido has recently said (maybe half-jokingly) that modules like `re` or `io` "have two-letter names for a reason", and should be flatly imported into another module (for example, new `typing`). And it is a fact that <pre class='brush: python'>import re is quite shorter than from re import findall x = re.findall(... x = findall(... </pre> But since you're not exporting `re` further, and the function is short, it doesn't really matter here.] <hr /> Now let's go to some more important stuff. Your regex is a bit misleading. First, there is really no need for '+': since you're doing a findall, and `''.join`ing at the end, it's simpler to just findall `'[A-Z]'`. Second, stylistically, it's better to write regexen as raw strings: `r'[A-Z]'`. Here it doesn't matter, but if you later add something else to the regex, it's possible you'll forget to `r` it. Python3.5 has finally deprecated the unrecognized backslashes (for example, interpreting `'\d'` as `'\d'`) and they will be removed in 3.6, so it will become riskier to write regexen as ordinary strings. And third, yes, the precondition effectively says that the input is ASCII, so you can get away with `[A-Z]`. But for a more serious applications, it's always better if the method works for the whole of Unicode. Unfortunately, Python's `re` module does not support "Unicode character properties" (`r'\p{Lu}'`), which is the right method to solve this kind of problem with a regex. (There is another module, `regex`, which is trying to better support this and many other features, and eventually replace `re` in Python's stdlib, but so far hasn't had much success - maybe because its name does not consist of just two letters.:)) Pythonic way of testing for an uppercaseness is the string method, `str.isupper`. Of course, then you don't need findall, nor regexen at all, but you need a `filter` builtin. With it you can write a better representation of what you really need: <pre class='brush: python'>x = filter(str.isupper, text) </pre> And the most important thing. Naming is very important in Python. There is nothing to be gained by picking a semantically empty name like `x`, and then reusing it for two very different things. Of course, when there are several steps, and intermediate results deserve semanticful names, there is value in writing things vertically: <pre class='brush: python'>uppercases = findall(r'[A-Z]', text) message = ''.join(uppercases) return message </pre> But if you don't consider intermediate results important enough to deserve meaningful names, then it's really better to just write the thing horizontally: <pre class='brush: python'>return ''.join(findall(r'[A-Z]', text)) </pre>

anonymorf on Sept. 1, 2015, 2:37 p.m.
Awesome! Thanks for your comprehensive answer.

andrewg12 on Sept. 5, 2015, 7:23 p.m.
What the heck is this thing!? I have a lot to learn

veky on Sept. 6, 2015, 8:34 a.m.
Of course. The question is, do you want to learn? If yes, it can be arranged. :-) For starter, try to read the comments. Or alternatively, try to ask a more coherent question than <blockquote> What the heck is this thing? </blockquote> (I can answer that, too, but maybe it would be more helpful if you asked a concrete thing.) It is a function that takes text, and gives a new string consisting of only uppercase letters from text. Concretely, filter picks only those elements of text where str.isupper gives true, and ''.join joins these elements into a string (with nothing, '', in between). Clearer?

andrewg12 on Sept. 7, 2015, 7:42 p.m.
just kinda wrote down my initial reaction and thoughts. Yes I want to learn, and I have been learning a lot from your solutions. I will check out the comments, but lambda looks inimidating, but I'll research, thanks.

veky on Sept. 8, 2015, 7:28 a.m.
The problem with lambdas, I think, is that people give them wrong intuition. Python has a quite strict divide between expressions and statements. When you see a bunch of statements repeated, it's natural to define a function with def, and then call it whenever you want to execute those statements. When you have an expression calculation repeated many times, or you just need it to have a specific name (like on CiO), you can get by with a def that just returns the expression (since function call is also an expression), but it's more natural to use lambda. <pre class='brush: python'> f = lambda t: t*3 + 1 a = x*3 + 1 ~~~> a = f(x) b = (-y*3 + 1 + z[1]*3 + 1)*3 + 1 b = f(f(-y) + f(z[1])) </pre> Of course, it's even better if the name `f` is semantically meaningful. :-)

brubru777 on Sept. 10, 2015, 12:56 p.m.
I was surprised to see so many negative comments for your solution. It is a similar to mine but a bit more compact and I dare say "elegant". Whether it respects the format you're supposed to use for answering or not isn't really relevant. The purpose of being able to look at other people's code is to learn by seeing different ways of coding. Your code taught me a lot. Thank you!

veky on Sept. 10, 2015, 4:25 p.m.
No problem. Thank you for your kind words.

suic on Sept. 10, 2015, 9:43 p.m.
After I've read a few long threads under your solutions I have to say, that lambdaphobia is a common phenomenon on CiO :)

veky on Sept. 11, 2015, 2:36 a.m.
Considering how long you follow me here, it took you surprisingly long to realize that. :-D But I won't give up. Many people have told me I have finally managed to demystify lambdas for them. (See <a href="http://www.checkio.org/mission/secret-message/publications/veky/python-3/filter/share/efd4e41ef30dcac9af99dcd6c84ae149/#comment-29982">this comment</a>.) It's too important an experience to be just given up on because of some people who think that Python is LISPifying, that PEP8 applies to hobby oneliners, or that "lambda is just a letter in the Greek alphabet" (this one is the best:)).

suic on Sept. 11, 2015, 12:35 p.m.
I know, it took quite a while :) Lambdaphobia came into my mind as a kind of epiphany and I had to write it down. Continue in your important mission of delispifying, antipep8cargocultism and lambdaphobia therapy :) The last one is a kind of never-ending story here.

amenokimy on Oct. 18, 2015, 12:21 p.m.
Hello again. Your code is so impressive. You are clearly smarter than me!

veky on Oct. 18, 2015, 1:35 p.m.
<blockquote> You are clearly smarter than me! </blockquote> Might be so, might be not. But you cannot conclude it just looking at the code. If someone saw you doing what you do best, they'd probably conclude you're incredibly smart. :-)

redvan on Nov. 15, 2015, 5:19 p.m.
Not bad

audrey_sean on Jan. 10, 2016, 1:06 p.m.
What's the difference between str.isupper and str.isupper( ) ?

tarun on Feb. 29, 2016, 8:39 a.m.
q

veky on Feb. 29, 2016, 11:43 a.m.
?

deamon32 on April 10, 2016, 6:30 a.m.
Cool, still need to wrap my head around the lambda usage a bit

veky on April 11, 2016, 5:42 a.m.
<pre class='brush: python'>def find_message(text): return ''.join(filter(str.isupper, text)) </pre> There. Just a mechanical transformation. Write a regex for homework. :-D

wuhuarui1992 on April 27, 2016, 8:36 p.m.
thanks veky. str.isupper is new to me. learned it.

ljy95135 on May 24, 2016, 8:52 a.m.
isupper and filter, I use isupper and list comprehension, the filter is new and nice for me.

veky on May 24, 2016, 10:11 a.m.
Listcomp is usually ok if you really do want a list as an output. Here you don't: you want str, so you aggregate with ''.join. That means you can just lose the [] and thing will work (and be clearer and faster and leaner): <pre class='brush: python'>''.join(character for character in text if character.isupper()) </pre> And then you see that the name "character" is not really semantically relevant very much, and it's quite long, and it is written three times above, so you start looking whether there is a better way. And there is: since .isupper() can be called as a predicate when unbound (str.isupper), you can just filter text by that predicate.

ljy95135 on May 24, 2016, 10:15 a.m.
Yes! In this situation, filter is even more natural. :)

veky on June 17, 2016, 9:27 p.m.
... which is exactly why I used it here. And named my solution by it. :)

james.t.hurst on Sept. 12, 2016, 8:23 p.m.
Really like the terseness.

veky on Sept. 12, 2016, 8:32 p.m.
Then you'll probably like <a href="https://py.checkio.org/mission/secret-message/publications/veky/python-27/those-were-the-days-my-friend/">this one</a> even more. ;-)

badwall on Sept. 20, 2016, 5:16 a.m.
I learnt from this code and the whole discussion around it (OMG, hot discussions about programming!) a lot more, then for the last two weeks of reading Python Beginner's textbook. Great work and many thanks. My solution to this problem is... well, is of the level of someone learning to code for about a month, but such brain stimulating solutions can really help a newbie to understand, that code differs, as well as approaches to different problems. It doesn't mean that I'll be able produce such tense code in next several months, but it really tells you that there is no boundaries in programming. Thanks for this!

veky on Sept. 20, 2016, 7:48 a.m.
That's why I love CiO. Yes, we're passionate about programming. And you ain't seen nothing yet. ;-D

saxo on Jan. 20, 2017, 5:11 p.m.
Hello, I'm just wondering why <pre class='brush: python'>filter(str.isupper, text) </pre> will work but not any other variable name ? eg : <pre class='brush: python'>filter(char.isupper, text) </pre> . I know str() is a method to convert something into a string, but I don't get it here ?

veky on Jan. 21, 2017, 6:04 a.m.
char doesn't work since it's an unknown name to Python. <pre class='brush: python'>>>> char NameError: name 'char' is not defined </pre> But `str` is not a method (it cannot be called as x.str(...) for some x, passing x as a first argument). You could say it's a function, but the real truth is something even more beautiful. ;-) <pre class='brush: python'>>>> str <class 'str'> </pre> str is a class. What you've seen is simply casting, using an object of some type to construct an object of another type while preserving as much of its value as possible. But that's only one usage of a class. You can check whether something is a str: <pre class='brush: python'>>>> isinstance(b"Ceci n'est pas un", str) False </pre> You can inherit from it: <pre class='brush: python'>>>> class Sentence(str): words = str.split >>> example = Sentence('Today is a beautiful day.') >>> for word in example.words(): print(word) </pre> And of course, you can access its methods and call them as functions. <pre class='brush: python'>>>> str.upper <method 'upper' of 'str' objects> >>> _('hm?') 'HM?' >>> _ == 'hm?'.upper() True </pre> I hope it's clearer now. If there are any more questions, feel free to ask.

saxo on Jan. 22, 2017, 4:09 p.m.
Thank you for your answer. It gave me a good basis for the understanding of those concepts ! Great help :)

marlonxteban on March 30, 2017, 4 a.m.
like it :-D

sharc06 on Sept. 5, 2017, 9:22 p.m.
Looks good. I enjoy one line coding when it makes sense and this makes sense.

flpo on Sept. 8, 2017, 11:45 p.m.
In (late) response to first comment. If this one should be in “Creative” category, what about <a href="https://py.checkio.org/mission/secret-message/publications/flpo/python-3/find_message-str_joinfind-message/">this one</a>? :-D

veky on Sept. 9, 2017, 8:26 a.m.
Then that one should be in str('WAT?!').staticmethod(dict=join)(['category']), of course. :-D Nice solution, BTW. :-D

alekseykonotop on Sept. 22, 2018, 9:41 a.m.
Cool!! You are pro!

veky on Sept. 22, 2018, 3:45 p.m.
I was, a long time ago. :-D

eminde on Oct. 25, 2018, 7:26 a.m.
this is awsome

DragonMachinations on Aug. 6, 2019, 3:08 p.m.
Thank you for posting this. This one reason I like this site. I managed to get my own solution as concise as I could make it, but then I managed to come here and learn three python principles. (lambda, join, and filter.) So thanks, it's an elegant solution.

veky on Aug. 6, 2019, 3:21 p.m.
Yes, that's one of main strenghts of CiO.

bsquare on Sept. 18, 2019, 1:27 p.m.
Nice, quick and clean solution!

undead2k on March 4, 2020, 6:13 p.m.
Amazing one liner!

Konstantin_Dubrovskij on June 13, 2021, 11:56 a.m.
this problem is not simple. it is elementary!

veky on June 13, 2021, 12:03 p.m.
I'm not quite sure what you are trying to say.

mikaeldovbnia on Feb. 16, 2023, 8:59 p.m.
This is the shortest solution I saw. I asked ChatGPT about how to shorten this code, but it said this is the shortest version.

veky on Feb. 17, 2023, 2:58 a.m.
ChatGPT is surprisingly not imaginative enough. Code could be shortened by switching to Python 2. :-D (See <a href="https://py.checkio.org/mission/secret-message/publications/veky/python-3/filter/share/efd4e41ef30dcac9af99dcd6c84ae149/#comment-22839">various other</a> approaches.)

mikaeldovbnia on Feb. 17, 2023, 3:06 a.m.
I didn't see that comment. I didn't know you could make it even more easy and short. The first solution in that comment is very short and good.

veky on Feb. 17, 2023, 10:08 a.m.
If you like my solutions, you should probably read my comments too, since there is a high probability you'll like them. :-)

mikaeldovbnia on Feb. 17, 2023, 3:07 a.m.
Also another question. Can I use functools module in python 3?

veky on Feb. 17, 2023, 10:08 a.m.
<blockquote> Can I use functools module in python 3? </blockquote> What makes you think you can't?

mikaeldovbnia on Feb. 17, 2023, 3:46 p.m.
Yeah, okay. Thanks for the information. I appreciate it. I never really studied functools with my tutor. I only finished itertools.

veky on Feb. 18, 2023, 3:33 a.m.
Ah, I think I see what happened. The reason why the solution partial(filter, str.isupper) doesn't work in Python 3 is that filter has different semantics in Python 2, not that functools is unavailable in Python 3.

mikaeldovbnia on Feb. 18, 2023, 2:50 p.m.
Yeah i tried this code and it always says <filter object at 0x10369d6c0>. I tried using the "*" sign (but momentarily realized I can't put i into a variable at the very start) and i tried using sum(, []), but didn't work.

veky on Feb. 26, 2023, 7:33 a.m.
I'm not sure what "put i into a variable" means. To join strings, you need ''.join (sum is explicitly disabled for strings, and * can't make a string since there are no string displays).

GitHub

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.