Using QwtSeriesData as data bridge for large datasets

Hi, I'm new to qt/qwt and am having some performance issues with large datasets. (linux OS, qt 5.7, qwt-6.1.3)
I've been looking at a few post about using QwtSeriesData as a data bridge for large datasets, thus having no need to copy the data into a QVector.
http://www.qtcentre.org/threads/2843...ng-application
http://www.qtcentre.org/threads/6729...-data?p=295423

My dataset consists of a large array of doubles and a static increment value, e.g.

Code:

---

int numPoints = 10000000
double data[numPoints]
double delta = .054;
QVector<QPointF> points

---

As of now I copy the data with an associated x value into a QVector<QPointF>

Code:

---

for (i=0; i<numPoints; i++)
{
   time = (i)*delta;
   points.append(QPointF(time, data[i]))
}

---

but unfortunately with my limited experience I'm unsure how to best reuse the linked examples for my application. Any basic examples to get me started would be much appreciated.

Code:

---

class YourData: public QwtSeriesData<QPointF>
{
    ....
    virtual size_t size() const
    {
        return 1000000;
    }
 
    virtual QPointF sample( size_t i ) const
    {
        return QPointF( i * delta, m_values[i] );
    }
 
    // optional
    virtual QRectF boundingRect() const
    {
        return QRect( 0.0, yMin, size() * delta, ymax - yMin );       
    }
 
private:
    double *m_values;
};

---

When workin with huge datasets you might want to have a look at the QwtPlotCurve::FilterPointsAggressive from svn trunk.

Uwe

Exactly what I needed. Thanks for the code snippet and the example recommendation Uwe