** UnhandledExceptionEventHandler :: OFFICIAL LC / CHATTER THREAD ** - Page 1038 - Computer Technical Help

Simple? Probably not, lol.

I'm pretty sure this can be solved with linear algebra, which, don't hold me to this, would make it around nlogn. As I understand it, there is no way to do matrix calculations any faster than that, but Linear Algebra is my worst math and matrices are my worst data structures.

The point of the sort is to create a seed of values, and prevent you from checking how each value works. The problem is that, no matter what you try, you can end up with some brute force at the end, given some combination of values.

Are we actually talking about the same problem?

https://codefights.com/challenge/Qjts7cukDvYpDW4Bc

What size would that matrix be anyway, and what exactly is n in your notation?

Going by the comments the best solutions seem to be O(a.len * (b.len-a.len)) runtime, I guess that's just some clever variation of the stuff i posted.

Quote

11-01-2016 , 02:48 PM

#25927

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

Cleaned it up a bit, this is the best i can do with that approach:

Spoiler:

This is O(a.len * (b.len-a.len)) runtime and O(b.len-a.len) memory.

Quote

11-01-2016 , 03:07 PM

#25928

daveT

S.A.G.E. Master

Join Date: Jun 2005 Posts: 23,955

I'm only considering B as N in the notation. The equation for the matrix would probably be M * B = C where size(M) = AxB. Once again, not my most awesome subject.

The runtime they call the best solution doesn't even make sense. If len(a) = len(b), then you have O(a * 0), which I'm pretty sure isn't a valid runtime.

Quote

11-01-2016 , 03:11 PM

#25929

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

If your matrix is size AxB then filling it is already higher complexity than the entire solution posted above, no?

Runtime is technically O(a * (b-a+1)) i guess

Spoiler:

Quote

11-01-2016 , 04:09 PM

#25930

daveT

S.A.G.E. Master

Join Date: Jun 2005 Posts: 23,955

The runtime, once again, makes no sense. Think of these situations:

a.len = b.len; a * (b - a + 1) = a
b.len = 1; a * (1 - a + 1) => do fuzzy math by dropping the constants => a * a.

The above is saying this:

a = [1]
b = [1, 2, 3, 4, 5, 6, 7, 8]

is slower than this:

a = [1, 2, 3, 4, 5, 6]
b = [1, 2, 3, 4, 5, 6, 7, 8]

I think this may be a variation on the Longest Common Sequence Problem: https://en.wikipedia.org/wiki/Longes...quence_problem

I never heard of LCS before, but I guess it is a major thing.

Another part of me thinks there is a bit of knapsack going on here as well. That can be optimized using bisection search.

Quote

11-01-2016 , 04:11 PM

#25931

blackize5

old hand

Join Date: Sep 2012 Posts: 1,358

Quote:

Originally Posted by clowntable

I'm drinking the Elixir coolaid and it tastes fantastic (currently working through the exercism.io stuff). Don't know why it took me so long. Pipe operator is elite (hi Closure folks) and I can finally use some of the old Prolog magic. The design of the web framework I'm learning simultaneously (Phoenix) aligns very well with the way I think so that's nice. Performance seems to be excellent as well...yay?!

I very much feel like this is it. New language for all side projects that don't strictly require another one (like when working with a game engine etc.).

I recently built an internal tool with Elixir/Phoenix and came away with similar impressions.

Really fun and fast language and framework. You can see the rails inspiration but it feels like the combination of the language and the framework force developers into better design

I ran into some trouble with authentication because it didn't seem like there was a fully featured library that isn't oauth, but the ecosystem seemed pretty great other than that.

As we build more microseconds I'm definitely going to push to implement in elixir where we can.

Quote

11-01-2016 , 04:23 PM

#25932

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

Quote:

Originally Posted by daveT

The above is saying this:

a = [1]
b = [1, 2, 3, 4, 5, 6, 7, 8]

is slower than this:

a = [1, 2, 3, 4, 5, 6]
b = [1, 2, 3, 4, 5, 6, 7, 8]

Huh?

Case A: 1*(8-1+1) = 9
Case B: 6 *(8-6+1) = 18

I'm a bit confused what we are arguing though, the code is posted right above and it's only a bunch of lines. Are you saying the code doesn't always find the correct answer, or that the runtime is different than stated?

Quote

11-01-2016 , 04:38 PM

#25933

daveT

S.A.G.E. Master

Join Date: Jun 2005 Posts: 23,955

The initial claim is this:

Quote:

Originally Posted by suzzer99

Interesting looking coding challenge that showed up on my FB feed with the clickbait title "94% of coders can't do this w/o using brute force":

Unless I'm grossly misunderstanding your solution... which apparently is a pattern over these past 2 days.

Quote

11-01-2016 , 04:48 PM

#25934

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

Yeah, this isn't a brute force solution to the problem. Brute forcing would be to generate all possible b' subsequences (that's choose(b, a) subsequences) and evaluating every single one.

Quote

11-01-2016 , 11:56 PM

#25935

suzzer99

Save the Cheerleader, Save the World

Join Date: Nov 2005 Posts: 99,984

I'm ****ing done with sublime. 45 minutes of trying to get eslint working, googling - weird answers, config files all over the place, global vs. local. This **** should just work like it does with Atom, Webstorm or VS Code.

Quote

11-02-2016 , 12:04 AM

#25936

Grue

Pooh-Bah

Join Date: Mar 2004 Posts: 5,667

wat? use gulp. I don't do sublime build anything, that's not what its good for.

Quote

11-02-2016 , 01:20 AM

#25937

suzzer99

Save the Cheerleader, Save the World

Join Date: Nov 2005 Posts: 99,984

Well now I'm back to sublime. No way to open multiple folders in one window in VS Code - total dealbreaker. https://visualstudio.uservoice.com/f...=1&per_page=20

Quote

11-02-2016 , 01:20 AM

#25938

suzzer99

Save the Cheerleader, Save the World

Join Date: Nov 2005 Posts: 99,984

Quote:

Originally Posted by Grue

wat? use gulp. I don't do sublime build anything, that's not what its good for.

Gulp is better for real-time linting/hinting as you type than Sublime?

Quote

11-02-2016 , 09:27 AM

#25939

Grue

Pooh-Bah

Join Date: Mar 2004 Posts: 5,667

I use sublime but don't use any its build stuff, I have gulp watch -> lint for on the fly linting with notifying. Simplified version:

Code:

let file;

gulp.task('watch', () => {
	gulp.watch(['./src/frontend-scripts/**/*.js*', './routes/**/*.js', './__test__/*.js'], e => {
		file = `./${e.path.split('/Users/xxx/repo/')[1]}`;
		gulp.start('lint');
	});
});

gulp.task('lint', () => {
	return gulp.src(file)
		.pipe(eslint())
		.pipe(plumber())
		.pipe(eslint.format())
		.pipe(eslint.failAfterError())
		.on('error', () => {
			notifier.notify({title: 'ESLint Error', message: ' '});
		});
});

Works pretty well. I also have a lint-all task which runs on startup.

Quote

11-02-2016 , 11:50 AM

#25940

suzzer99

Save the Cheerleader, Save the World

Join Date: Nov 2005 Posts: 99,984

That's not the same as in-editor linting though.

Quote

11-02-2016 , 01:15 PM

#25941

Grue

Pooh-Bah

Join Date: Mar 2004 Posts: 5,667

fair enough but using node-notify it pops up when there's issues and I just glance at other screen

Quote

11-02-2016 , 02:09 PM

#25942

suzzer99

Save the Cheerleader, Save the World

Join Date: Nov 2005 Posts: 99,984

Still not the same, especially when 90% of the time you're working on a laptop with a 13% screen.

Now my sublime jshint is broken after all the ****ing around with es lint. AWESOME. Once I get this working I'm never touching it again. If I need es lint I will use a different editor.

Quote

11-02-2016 , 02:24 PM

#25943

n00b590

adept

Join Date: Jun 2010 Posts: 978

Quote:

Originally Posted by suzzer99

Still not the same, especially when 90% of the time you're working on a laptop with a 13% screen.

noob

Quote

11-02-2016 , 02:32 PM

#25944

suzzer99

Save the Cheerleader, Save the World

Join Date: Nov 2005 Posts: 99,984

I have two thunderbolts on my desk in my bedroom. I just prefer to hang out in the living room, or in a coffee shop - unless I really need the screen real estate for some task.

Quote

11-02-2016 , 03:15 PM

#25945

Grue

Pooh-Bah

Join Date: Mar 2004 Posts: 5,667

Non ultrawide peasants ITT.

Quote

11-02-2016 , 03:30 PM

#25946

RustyBrooks

Carpal \'Tunnel

Join Date: Feb 2006 Posts: 24,647

I'm picturing a 13% screen as a screen where it's like a giant bezel with a tiny monitor. Like those toy kid computers they used to make (maybe still do for all I know)

Quote

11-02-2016 , 04:27 PM

#25947

suzzer99

Save the Cheerleader, Save the World

Join Date: Nov 2005 Posts: 99,984

Heh - 13" obviously.

Quote

11-02-2016 , 05:31 PM

#25948

codeartisan

centurion

Join Date: Jun 2008 Posts: 181

Quote:

Originally Posted by Grue

Non ultrawide peasants ITT.

80 chars FTW

Quote

11-02-2016 , 09:52 PM

#25949

RogerKwok

grinder

Join Date: Apr 2009 Posts: 422

esh no IDE for js in 2016? do you type out your imports you pleb

Quote

11-02-2016 , 10:18 PM

#25950

Ryanb9

Carpal \'Tunnel

Join Date: Aug 2006 Posts: 7,459

need some quick help. im using juypter (iPython) and i have text file that i used re to split into words, they are all in a list now. i need to come up with a data frame (i guess?) where it has each word and the corresponding frequency. thing is i cant use iteration. i can use numpy or pandas to do iteration for me, but i cant do iteration myself.

im thinking about using "apply" but wtf i have no idea what im doing... i hate python. i suck at it. any ideas? lol

Quote

Page 1038 of 1603

First

38 538 838 938 988 1018 1028 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1048 1058 1088 1138 1238 1538

Last

Post Reply Subscribe

...

Page 1038 of 1603

First

38 538 838 938 988 1018 1028 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1048 1058 1088 1138 1238 1538

Last