digitalmars.D.learn - Reading array of integers readln performance issues
- kerdemdemir (22/22) Jun 11 2015 Hi;
- Dennis Ritchie (7/29) Jun 11 2015 Your algorithm works for about quadratic time. For N = 200000
- anonymous (17/25) Jun 11 2015 `std.array.split` is eager. It may be faster if you use the lazy
- kerdemdemir (5/5) Jun 12 2015 Thanks a lot for your great advices and exaamples. Yes if I don't
Hi; To learn D better and challanging myself I am tring code computation's with D. There is a question which is about reading a line of integer which consist of 200000 elements. My solution fails because "Time limit exceeded", I thought it is because of my algorithm first. I realize time limit is exceeded even before my algorithm starts while reading line of integers. I understand this by giving a wrong answer to question after readln statement. I did that to get a "wrong answer error" but my code still get a "Time limit exceed" error because "readln" takes very long time. Can I achieve something faster than code below? auto peopleMoney = stdin.readln().split().map!(a => to!int(a)).array(); if (peopleMoney.length == 200000) writeln(":("); Regards Erdem Ps: I do not want to bore you with long code, but I am sending link to whole program anyway if anyone need. http://codeforces.com/contest/549/submission/11537206
Jun 11 2015
On Thursday, 11 June 2015 at 19:56:00 UTC, kerdemdemir wrote:Hi; To learn D better and challanging myself I am tring code computation's with D. There is a question which is about reading a line of integer which consist of 200000 elements. My solution fails because "Time limit exceeded", I thought it is because of my algorithm first. I realize time limit is exceeded even before my algorithm starts while reading line of integers. I understand this by giving a wrong answer to question after readln statement. I did that to get a "wrong answer error" but my code still get a "Time limit exceed" error because "readln" takes very long time. Can I achieve something faster than code below? auto peopleMoney = stdin.readln().split().map!(a => to!int(a)).array(); if (peopleMoney.length == 200000) writeln(":("); Regards Erdem Ps: I do not want to bore you with long code, but I am sending link to whole program anyway if anyone need. http://codeforces.com/contest/549/submission/11537206Your algorithm works for about quadratic time. For N = 200000 your algorithm will work terribly long. The function `readln()` nothing to do with: http://codeforces.com/contest/549/submission/11476513 A faster way of `scanf`, but it will not help you, because your algorithm is slow.
Jun 11 2015
On Thursday, 11 June 2015 at 19:56:00 UTC, kerdemdemir wrote:Can I achieve something faster than code below? auto peopleMoney = stdin.readln().split().map!(a => to!int(a)).array(); if (peopleMoney.length == 200000) writeln(":(");`std.array.split` is eager. It may be faster if you use the lazy `std.algorithm.splitter`: auto peopleMoney = stdin.readln().splitter().map!(to!int).array(); [...]Ps: I do not want to bore you with long code, but I am sending link to whole program anyway if anyone need. http://codeforces.com/contest/549/submission/11537206Seeing that you get the number of elements beforehand, you can preallocate the array, avoiding relocations of the data as elements are appended: peopleMoney = new int[peopleCount]; copy(stdin.readln().splitter().map!(to!int), peopleMoney); But these are tweaks. They may improve performance a little, but it won't be drastic. And anyway, I find it hard to believe that the original version takes more than a second to parse the input. Maybe try returning from the function in addition to printing ":(". Otherwise the program goes on, and the site may not show the output if the program as a whole took too long.
Jun 11 2015
Thanks a lot for your great advices and exaamples. Yes if I don't return; web-site won't show it as "wrong answer". As a learner I am very happy with the responsiveness of the community. Regards
Jun 12 2015