is this typedef valid?

[lou zion]

hi all,

i want to create a vector where each element is an array of 40 doubles. is
this valid? something like:

void abc( std::vector<double> InVals)
{
typedef double DataSeriesType[40];
std::vector<DataSeriesType> DataSeriesX;

DataSeriesType NewData;
DataSeriesX.push_back(NewData);

for (int i=0; i<(int)InVals.size(); i++) {
DataSeriesX[0][i]=InVals[i];
}
}

the above is not good code, i know, i'm just using it as an example of
syntax, not style.

essentially i want to store an array (or possibly a vector) in a vector. if
i just store a pointer to an array in a vector, what happens if i delete
that element? do i need to deallocate storage somehow? i'd prefer an array,
but if there's a better way to do this, please enlighten.

thnx :)

lou

[Ioannis Vranos]

lou zion wrote:

hi all,

i want to create a vector where each element is an array of 40 doubles. is
this valid? something like:

void abc( std::vector<double> InVals)
{
typedef double DataSeriesType[40];
std::vector<DataSeriesType> DataSeriesX;

DataSeriesType NewData;
DataSeriesX.push_back(NewData);

for (int i=0; i<(int)InVals.size(); i++) {
DataSeriesX[0][i]=InVals[i];
}
}

the above is not good code, i know, i'm just using it as an example of
syntax, not style.

essentially i want to store an array (or possibly a vector) in a vector. if
i just store a pointer to an array in a vector, what happens if i delete
that element? do i need to deallocate storage somehow? i'd prefer an array,
but if there's a better way to do this, please enlighten.

#include <vector>

void abc(const std::vector<double> &InVals)
{
using namespace std;

vector<vector<double> >DataSeriesX;

vector<double> NewData(40);

DataSeriesX.push_back(NewData);

for(vector<double>::size_type i=0; i<InVals.size(); ++i)
DataSeriesX[0][i]=InVals[i];
}
or
#include <vector>

void abc(const std::vector<double> &InVals)
{
using namespace std;

typedef vector<double>DataSeriesType;

vector<DataSeriesType> DataSeriesX;

DataSeriesType NewData(40);

DataSeriesX.push_back(NewData);

for(vector<double>::size_type i=0; i<InVals.size(); ++i)
DataSeriesX[0][i]=InVals[i];
}


--
Ioannis Vranos

http://www23.brinkster.com/noicys

[Victor Bazarov]

"lou zion" <il********@adelphia.net> wrote...

i want to create a vector where each element is an array of 40 doubles. is
this valid?
No. Arrays cannot be stored in a standard container because they do not
satisfy the requirements. The easiest thing is to wrap them in a struct.
something like:

void abc( std::vector<double> InVals)
{
typedef double DataSeriesType[40];
std::vector<DataSeriesType> DataSeriesX;

DataSeriesType NewData;
DataSeriesX.push_back(NewData);

for (int i=0; i<(int)InVals.size(); i++) {
DataSeriesX[0][i]=InVals[i];
}
}

the above is not good code, i know, i'm just using it as an example of
syntax, not style.

essentially i want to store an array (or possibly a vector) in a vector.
if i just store a pointer to an array in a vector, what happens if i
delete that element? do i need to deallocate storage somehow? i'd prefer
an array, but if there's a better way to do this, please enlighten.

If you store a pointer to a dynamic array that you get from using 'new[]',
then just before deleting that element from the vector you will need to
deallocate the memory. A bit of a hassle. Wrapping the array in a struct
is easier.

V

[lou zion]

"Victor Bazarov" <v.********@comAcast.net> wrote in message
news:u6********************@comcast.com...

"lou zion" <il********@adelphia.net> wrote...

i want to create a vector where each element is an array of 40 doubles.
is this valid?

No. Arrays cannot be stored in a standard container because they do not
satisfy the requirements. The easiest thing is to wrap them in a struct.

hi, thanks to both of you. these'll work well.

lou

[E. Robert Tisdale]

Ioannis Vranos wrote:

lou zion wrote:

I want to create a vector
where each element is an array of 40 doubles.

#include <vector>

void abc(const std::vector<double> &InVals)
{
using namespace std;

vector<vector<double> >DataSeriesX;

vector<double> NewData(40);

DataSeriesX.push_back(NewData);

for(vector<double>::size_type i=0; i<InVals.size(); ++i)
DataSeriesX[0][i]=InVals[i];
}

The problem with this solution is that
an object of type vector<vector<double> >
is *not* the object that Lou Zion describes.
A vector<double> has a *flexible* extent
whereas double[40] has a *rigid* extent.

There is also no range checking for 40 < InVals.size().

[Ioannis Vranos]

E. Robert Tisdale wrote:

The problem with this solution is that
an object of type vector<vector<double> >
is *not* the object that Lou Zion describes.

Actually his words were:

"essentially i want to store an array (or possibly a vector) in a vector."

A vector<double> has a *flexible* extent
whereas double[40] has a *rigid* extent.

What do you mean. That the amount of its contents does not change? Well,
if someone wants to preserve the amount of contents, then he should not
change them.

There is also no range checking for 40 < InVals.size().

?


--
Ioannis Vranos

http://www23.brinkster.com/noicys