Shortest speedy solution for The Most Wanted Letter by veky - python coding challenges

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

Shortest solution in Speedy category for The Most Wanted Letter by veky

checkio=lambda t:max(map(chr,range(97,123)),key=t.lower().count)

Sept. 26, 2013

Comments:

oduvan on Oct. 10, 2013, 12:13 p.m.
"Harder, Better, Faster, Stronger" - Daft Punk "Harder, smaller, shorter" - Veky :)

CyroAvernus on Nov. 3, 2013, 1:30 p.m.
Really the shortest and nice usage of the map function.....

veky on Nov. 3, 2013, 4:04 p.m.
When I say "shortest", it's really shortest. Kolmogorov FTW. :-D

StefanPochmann on April 14, 2015, 3:31 p.m.
Your Kolmogorov is wrong: <a href="http://www.checkio.org/mission/most-wanted-letter/publications/StefanPochmann/python-3/shortest-p/">My shorter solution</a>

veky on April 15, 2015, 6:29 a.m.
It just depends on how you define "solution". If you define it just as "passes all tests", you're right, but that's unstable - CiO has been known to add tests to already existing tasks. I define a solution as something that solves the given task, not taking into account what tests are. That's why preconditions are there, after all.

StefanPochmann on April 15, 2015, 5:45 p.m.
Yeah, just a joke. My real solution was exactly the same as yours. I tried to improve it, but failed. Then I saw your mention of Kolmogorov, which makes me think of trying programs of increasing size until one happens to do the job. Yeah I know that that doesn't work, but it makes me think of that anyway and it's kinda what I did to find my string.

veky on April 16, 2015, 10:13 a.m.
"kinda" is deceptive. You added characters to the string, not instructions to a program. I know that "code is data is code" is a popular belief among LISP-fanatics, but not going into whether that belief is right or wrong, it is counterproductive in Python philosophy. Python doesn't have macros, tail call elimination, switch statement, or any of these wonderful things that blur the distinction.

A_bak_cfff596eda7840afb01132b2 on Nov. 25, 2013, 6:23 p.m.
Also it's the fastest one. I tried to compare this solution to other "Editor choices", and this one is almost twice as fast as the next one. (Python 2.7.3, average time of 100 cycles for 1000 iterations, with all the asserts for this task). Great job, veky!

veky on Nov. 25, 2013, 9:31 p.m.
Thanks. As I said many times, the less you write, the less the interpreter has to chow on. And it's frequently the deciding factor. But my motivation is never primarily speed. It's just a nice sideeffect. ;-)

rolle_bolle on Dec. 3, 2013, 3:52 a.m.
This is awesome! Nice work. By any chance, you would mind explaining to me why max(...) works instead of min(...)? I don't quite get that part. Very impressive... and mind blowing!

veky on Dec. 3, 2013, 8:34 a.m.
"key" is the key to the explanation. max is not by intrinsic order, but by key. That is, a is max of s with key k, if all(k(a)>=k(c) for c in s). If there are more elements with the same maximal key, max picks the first one (again not by intrinsic order, but by order in which they appear in s). Clearer?

rolle_bolle on Dec. 3, 2013, 10:59 p.m.
It sure is clearer. Thanks for taking the time to respond. I studied your answer quite a bit last night trying to wrap my head around it. It's awesome!

veky on Dec. 4, 2013, 12:38 a.m.
Thanks. Be sure to check other solutions of mine here (for other long winter nights:). ;-)

mr.floppy on Oct. 13, 2014, 2:53 a.m.
But don't try them for an insomnia cure: they're too short & far too interesting.

veky on Oct. 13, 2014, 6:46 a.m.
LOL. This is one of reasons I like CiO so much: those compliments I get here, that are often also too short and far too interesting. ;-D

serhiideliiev on Nov. 10, 2014, 11:56 p.m.
It couldn't be shorter.

veky on Nov. 11, 2014, 6:50 a.m.
I surely hope so... but Kolmogorov might surprise you sometimes. ;-)

caynan on March 13, 2015, 5:13 p.m.
There is any advantage of using <pre class='brush: python'>map(chr,range(97,123)) </pre> over using, apart from not having to import something? <pre class='brush: python'>string.ascii_lowercase </pre> Sorry for the stupid question.

veky on March 13, 2015, 11:03 p.m.
It's not stupid. This is just shorter (when you count in the import). And since here I was trying to golf, it seemed appropriate. You have another solution of mine, "key", where I use exactly that approach.

taras-sereda on April 2, 2015, 6:36 p.m.
Nice golf! I've just posted my new solution, pretending that code is short and speed. but when I opened yours, I was impressed again)) Good trick! <pre class='brush: python'>key=t.lower().count </pre>

veky on April 2, 2015, 7:36 p.m.
It's not a trick, it's just a syntax feature: . has greater precedence than (), so a.f(x) is not a.(f(x)), but (a.f)(x). Bound methods are functions.

taras-sereda on April 2, 2015, 7:51 p.m.
Thanks for pointing on syntax features. I understood this in the following way. t.lower() - returns lowered string and after that count applied which in turn will be used by key.

veky on April 3, 2015, 4:33 a.m.
t.lower() is a lowered t. But count is not "applied" (evaluated at some particular letter) until max starts keying the letters. The whole point is that the lookup of the function name is completely separable from evaluating it. :-) <a href="https://docs.python.org/3.5/tutorial/classes.html">https://docs.python.org/3.5/tutorial/classes.html</a>#method-objects <blockquote> Usually, a method is called right after it is bound [...] However, it is not necessary to call a method right away: `x.f` is a method object, and can be stored away and called at a later time. </blockquote>

taras-sereda on April 3, 2015, 8:01 a.m.
Thanks a lot! It means that in this case <pre class='brush: python'>t.lower().count </pre> is a method object.Right? Which will be called and evaluated latter. And this method object created by packing with pointers to instance object ( t.lower()) and function object (count). Which in turn will be called like this at evaluation time. <pre class='brush: python'>str.count(t.lower(),'a') </pre>

veky on April 3, 2015, 8:53 a.m.
Two terminology issues: "pointers" are better called "references", or even "attributes" here. And function object is in fact `str.count`. And it is really called as <pre class='brush: python'>t.lower().count('a') </pre> which is theoretically not the same (for example, if `t.lower()` is an instance of a subclass of `str`:). But of course this is just a nitpick.

taras-sereda on April 4, 2015, 1:26 p.m.
Nevertheless thanks for nitpicks :) It helps to solidify knowledge.

Flying_Eye on May 4, 2016, 10:17 p.m.
str.count complexity is linear, so complexity of this solution is (123 - 97) * len(t). Can we do better reading the string only once?

veky on May 5, 2016, 5:53 a.m.
<a href="https://checkio.org/mission/most-wanted-letter/publications/veky/python-3/1pass/">Yes</a>, but it's not so elegant. :-)

Q-Tang on Oct. 23, 2018, 2:01 a.m.
How does this work without def

veky on Oct. 23, 2018, 2:41 p.m.
lambda is an expression cousin of def. <pre class='brush: python'>function = lambda whatever: expression <~~~> def function(whatever): return expression </pre>

Q-Tang on Oct. 25, 2018, 12:12 a.m.
ohhhhh ok

Q-Tang on Oct. 25, 2018, 12:13 a.m.
ohhhhhh ok

roman.alilov on Sept. 4, 2019, 2:58 p.m.
Elegant

bsquare on Sept. 18, 2019, 10:22 a.m.
Very nice solution ... and you get nice congratz from others (Veky Punk ^^).

veky on Sept. 18, 2019, 10:49 a.m.
Ha? =_=

bsquare on Sept. 18, 2019, 12:08 p.m.
@oduvan once said: <blockquote> “Harder, Better, Faster, Stronger” - Daft Punk “Harder, smaller, shorter” - Veky :) </blockquote> ^^

veky on Sept. 18, 2019, 12:12 p.m.
Ah yes, Land of Weird Compliments. :-D

Strings and Integers

0 %

PyCharm The Python IDE for Professional Developers. Enjoy productive Python, web and scientific development with PyCharm. Download it now!