a map of queues? c++ STL

nkoplm · 07-30-2011, 03:41 AM

Hello, I am trying to store the data in my program in a map of priority_queues an I am having a bit of trouble doing it.

basically I am storing activities (ie watching tv, eating dinner, playing ping-pong, etc) and I want these event grouped by the day that they happened on.

my first thought was to use a multimap where the key is the date, and the value is the activity, however I would like to keep the activities in a specific order.

I have tried to find the proper syntax for this but have frustratingly come up empty. Even google provided little assistance.

could anyone here point me in the right direction?

thanks!

paulsm4 · 07-30-2011, 04:05 AM

Why not just:

1. Create a class that has your queue and your activity name
2. Create a map containing instances of this class?

nkoplm · 07-30-2011, 04:23 AM

Although that would work, I don't believe it has any advantage to my current design. The idea is to group activities together by which day thy were performed on, not by the type of activity that was performed. and in that respect, a "day" object would server no purpose except to encapsulate a queue. This would be extra complexity that is unneeded.

Though I could technically just have one big chrolological list of all the activites that have ever been performed, however, because the list could potentially get quite large, I need to consider running time and efficiency constraints. being able to break the list up into days, or months, or whatever, could greatly improve performance in some use cases.

If i have misunderstood you, or if i could clarify my problem better, let me know.
thanks.

SigTerm · 07-30-2011, 04:40 AM

Quote:

Originally Posted by nkoplm

Hello, I am trying to store the data in my program in a map of priority_queues an I am having a bit of trouble doing it.

basically I am storing activities (ie watching tv, eating dinner, playing ping-pong, etc) and I want these event grouped by the day that they happened on.

my first thought was to use a multimap where the key is the date, and the value is the activity, however I would like to keep the activities in a specific order.

I have tried to find the proper syntax for this but have frustratingly come up empty. Even google provided little assistance.

could anyone here point me in the right direction?

thanks!

Code:

std::map<Date, std::queue<Activity>>

or (recommended)

Code:

typedef std::queue<Activity> ActivityList;
typedef std::map<Date, ActivityList> ActivityMap;

Where Date and Actviity are your types. Keep in mind, that you should only be using references with element stored within map,

Code:

ActivityMap activityMap;
//initialization
Date currentDate = ...;
ActivityList &currentActivity = activityMap[currentDate];

otherwise you'll invoke copy constructor and won't be operating on element stored within map.

ta0kira · 07-30-2011, 11:08 AM

Quote:

Originally Posted by SigTerm

Code:

std::map<Date, std::queue<Activity>>

Does this compile for you? It shouldn't; it should be > >. I've seen this in other code and I'm wondering if some compilers allow it. Aside from that, I agree with your recommendation.
Kevin Barry

SigTerm · 07-30-2011, 11:59 AM

Quote:

Originally Posted by ta0kira

Does this compile for you?

It does:

Code:

#include <map>
#include <queue>

int main(int argc, char** argv){
	std::map<int, std::queue<int>> a;
	return 0;
}

Quote:

Originally Posted by ta0kira

It shouldn't;

Can you quote C++ standard saying so? (I understand that ">>" is technically a shift operator token).

Quote:

Originally Posted by ta0kira

I've seen this in other code and I'm wondering if some compilers allow it.

Microsoft Compiler allows it. g++ doesn't.

ta0kira · 07-30-2011, 03:44 PM

Quote:

Originally Posted by SigTerm

Can you quote C++ standard saying so? (I understand that ">>" is technically a shift operator token).

More importantly, can you quote the standard saying that >> is ever other-than an operator? Does this compile for you in MS?

Code:

#include <vector>

template <int Value>
struct constant
{ static const int value = Value; };

const int index = 1;

int main()
{
  std::vector <constant <1024 >> index> > instance;
}

The code in red would be a complete and valid definition statement if >> denoted the end of nested template instantiations, making the code in blue an error.

Note that a 2011 draft of the standard (14.1.1, 14.2.3) allows the syntax you used; however, it's not a part of the 2003 standard. That means you shouldn't expect it to compile with compilers that don't yet implement the 2011 standard (C++0x). Not to mention it will cause anyone using g++ problems.
Kevin Barry

edit: Also note that my example is invalid for C++0x but valid for C++03 (see 14.2.3).

SigTerm · 07-31-2011, 04:50 AM

Quote:

Originally Posted by ta0kira

More importantly, can you quote the standard saying that >> is ever other-than an operator? Does this compile for you in MS?

It doesn't unless put "1024 >> index" into () brackets.

Quote:

Originally Posted by ta0kira

That means you shouldn't expect it to compile with compilers that don't yet implement the 2011 standard (C++0x). Not to mention it will cause anyone using g++ problems.

I have not asked for assistance, so please refrain from telling me what to do or expect. Errors like this are quickly found during the first attempt to compile on different compiler, and are easy to deal with. Another thing is that I prefer to use typedefs for readability so it is unlikely that I'll run into this situation.

ta0kira · 07-31-2011, 10:57 AM

Quote:

Originally Posted by SigTerm

I have not asked for assistance, so please refrain from telling me what to do or expect. Errors like this are quickly found during the first attempt to compile on different compiler, and are easy to deal with. Another thing is that I prefer to use typedefs for readability so it is unlikely that I'll run into this situation.

Noted, but I wasn't providing you with assistance. I was more concerned with you providing a solution that won't compile on g++; such an error might not be obvious to a beginner, especially if it comes recommended by someone with experience.
Kevin Barry

nkoplm · 07-31-2011, 07:23 PM

Quote:

Originally Posted by ta0kira

I was more concerned with you providing a solution that won't compile on g++; such an error might not be obvious to a beginner, especially if it comes recommended by someone with experience.
Kevin Barry

Agreed. Thank you both for your help. I have it working now.