Jump to content

  • Log In with Google      Sign In   
  • Create Account


How to write a good function printing float matrix to console?


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
8 replies to this topic

#1 Misery   Members   -  Reputation: 305

Like
0Likes
Like

Posted 26 December 2012 - 05:31 PM

Hi there,

I am working on a matrix library. At the moment I am trying to write a function that will display the given matrix in the console.
However I don't really have any good idea how to get started. I mean, which stream should I use (for example iostream or stringstream), how to make formatting in details?

The main requirements are:
1) Display the numbers in given precision in fixed format, but if the length of resulting output in fixed format is too big, than convert (only this matrix element) to scientific format. (therefore I hesitate if I should use stringstream, as it easily alllows to check the length of the resulting output)

2) There should exist easy way of redirecting the stream to file.

3) It should be possibly efficient, however this is not the main goal

And other thing: should I buffer the output and then print the whole thing, or better to cout element after element. Won't be using stringstream for formatting, and then using cout for printing too slow?

I really need some general directions (but code snippets are very welcome too).

Thanks in advance,
Misery

Edited by Misery, 26 December 2012 - 05:33 PM.


Sponsor:

#2 SiCrane   Moderators   -  Reputation: 9432

Like
1Likes
Like

Posted 26 December 2012 - 05:48 PM

If you want to make your output function have a flexible choice of destination, you can make it take a std::ostream & argument for destination. Then whoever uses it can dump it to a std::stringstream, std::cout or a std::fstream. One option is to just overload std::ostream & operator<<(std::ostream &, const YourClass &) so you can use the formatted output operator.

#3 santa01   Members   -  Reputation: 307

Like
0Likes
Like

Posted 26 December 2012 - 06:16 PM

If you want to make your output function have a flexible choice of destination, you can make it take a std::ostream & argument for destination. Then whoever uses it can dump it to a std::stringstream, std::cout or a std::fstream. One option is to just overload std::ostream & operator<<(std::ostream &, const YourClass &) so you can use the formatted output operator.

 

This. And for requirement (1) I'd wrap the sprintf() call to make a fancy formatting (sorry for non C++ approach, but I don't know any decent STL classes to solve this issue).



#4 Bacterius   Crossbones+   -  Reputation: 8191

Like
2Likes
Like

Posted 27 December 2012 - 03:11 AM

I'm not sure why you worry about performance here. Writing to an I/O stream and displaying stuff on a terminal will always be slow, no matter how efficient your code is. Clearly this is a debugging feature, right? Also, do you want a human-friendly output, like:

4.888 3.901 1.213 9.885
0.818 1.831 4.918 5.192
3.192 9.091 1.119 2.283
7.004 1.193 1.717 4.483

Or a parser-friendly output, such as:

 

((4.888, 3.901, 1.213, 9.885), (0.818, 1.831, 4.918, 5.192), (3.192, 9.091, 1.119, 2.283), (7.004, 1.193, 1.717, 4.483))

The problem with the human-friendly output is that it spans multiple lines, which doesn't quite fit the concept of string (sure, you can use newlines, but you need to add them everywhere to make sure the matrix doesn't come out garbled because it started at the end of an existing line, etc...), whereas with the parser-friendly output, you can print it easily to the console or a file, possibly with a debug label, and then feed them to Mathematica or Wolfram|Alpha or even LaTeX (perhaps using a script) which'll display type-setted matrices to you (and you can even then perform computations on them to check them), but of course this can be less practical depending on your needs.


The slowsort algorithm is a perfect illustration of the multiply and surrender paradigm, which is perhaps the single most important paradigm in the development of reluctant algorithms. The basic multiply and surrender strategy consists in replacing the problem at hand by two or more subproblems, each slightly simpler than the original, and continue multiplying subproblems and subsubproblems recursively in this fashion as long as possible. At some point the subproblems will all become so simple that their solution can no longer be postponed, and we will have to surrender. Experience shows that, in most cases, by the time this point is reached the total work will be substantially higher than what could have been wasted by a more direct approach.

 

- Pessimal Algorithms and Simplexity Analysis


#5 Misery   Members   -  Reputation: 305

Like
0Likes
Like

Posted 27 December 2012 - 05:59 AM

@Bacterius: I need a user friendly output. I was just wondering if I can make my function better than it is now.



#6 MarkS   Prime Members   -  Reputation: 875

Like
0Likes
Like

Posted 27 December 2012 - 10:50 AM

If you want to make your output function have a flexible choice of destination, you can make it take a std::ostream & argument for destination. Then whoever uses it can dump it to a std::stringstream, std::cout or a std::fstream. One option is to just overload std::ostream & operator<<(std::ostream &, const YourClass &) so you can use the formatted output operator.

 

This. And for requirement (1) I'd wrap the sprintf() call to make a fancy formatting (sorry for non C++ approach, but I don't know any decent STL classes to solve this issue).

I'm curious why this was down-voted? There are no C++ functions that provide formatting like the old C functions. This has led me on many occasions to prefer the C functions over C++ or to mix them. If the person that down-voted this has a C++ solution, please post it.



#7 SiCrane   Moderators   -  Reputation: 9432

Like
1Likes
Like

Posted 27 December 2012 - 11:23 AM

There are no C++ functions that provide formatting like the old C functions.
Can you give an example of formatting that you can do with the C standard library functions that you can't do with the C++ standard library functions? I can't think of any (though I will freely admit that the C++ versions are much, much more annoying to use).

#8 MarkS   Prime Members   -  Reputation: 875

Like
0Likes
Like

Posted 27 December 2012 - 11:40 AM

You can, for example, set the number of digits to be printed after the decimal in a floating point number. You can set the number of leading/trailing zeros in a number. You can print anything in hexadecimal without writing a conversion function. You can choose to print in decimal or scientific notation. The list goes on. I've searched, and I cannot find this ability in any C++ stream. Maybe I haven't searched well enough?

 

[edit]

 

Well, what do you know...

 

http://www.cplusplus.com/reference/ios/ios_base/fmtflags/

 

How did I miss this?


Edited by MarkS, 27 December 2012 - 11:42 AM.


#9 Ohforf sake   Members   -  Reputation: 1576

Like
0Likes
Like

Posted 28 December 2012 - 12:54 AM

I usually like my matrices and vectors in Matlab-Syntax, such that I can just copy & paste them into Matlab/Octave. I feel that if the matrix is simple, I can look at it in the debugger, and if it is complicated, then actually looking at it won't do any good and further computations in Matlab/Octave are necessary anyways.

 

I know, this is rather unrelated to you question, but I thought I should suggest it, if only as an extra option.






Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS