Frequently Asked Questions QCChart2D for Java

FAQs

Is theQCChart2D for Java software backward compatible with the Charting Tools for Windows and the Graphics Class Librariesfor MFC ?
How do you create a chart with multiple coordinate systems and axes?
Can I add new axes, text objects, plot objects, and images to a chart after it is already displayed; or must I create a new chart from scratch taking into account the additional objects?
How do you zoom charts that use multiple coordinate systems?
How do you select a chart object and create a dialog panel that permits editing of that objects properties?
How do you handle missing data points in a chart?
How do you update a chart in real-time?
How do I prevent flicker when updating my charts on real-time?
How do you implement drill down, or data tool tips in a chart?
I do not want to my graph to auto-scale. How do I setup the graph axes for a specific range?
How do I update my data, and auto-rescale the chart scales and axes to reflect the new data, after it has already been drawn?
When I use the auto-scale and auto-axis routines my semi-log chart has the logarithmic axis scaled using powers of 10 (1, 10,100, 1000, etc.) as the starting and ending values, or as the major tick interval for labeling. How do I make my log graphs start at 20 and end at 50,000, with major tick marks at 20, 200, 2000 and 20000?
How do I create and use custom, multi-line string labels as the axis labels for my graph?
How do I place more than one graph in a view?
How do I use your software to generate GIF files?
Sometimes the major tick marks of an axis are missing the associated tick mark label ?
How do I change the order the chart objects are drawn? For example, I want one of my grid objects to be drawn under the charts line plot objects, and another grid object to be drawn top of the charts line plot objects.
How to I use a scrollbar object to control horizontal scrolling of the data in my chart?
I am trying to plot 100,000,000 data points and it takes too long to draw the graph. What is wrong with the software and what can I do to make it faster?
How do I get data from my database into a chart?
How do I use this charting software to generate chart images “on-the-fly”?

22. Can QCChart2D for Java be used to create programs and controls that run in web browsers?

Is the QCChart2D for Java software compatible with the Charting Tools for Windows and the Graphics Class Libraries for MFC ?

No, the QCChart2D for Java software is not compatible with the Quinn-Curtis Charting Tools for Windows (WIN-XXX-100), or Graphics Class Libraries for MFC (WIN-MFC-210). Instead it is a translation of QCChart2D for .Net product. Programs written using the .Net and Java versions will be very similar. You should have no problems recreating any charts that you created using our older Windows software; in most cases it will take far fewer lines of code.

How do you create a chart with multiple coordinate systems and axes?

A chart can have as many coordinate systems and axes as you want. A single coordinate system can have one or more x- and/or y-axes. The most common use for multiple axes in a single coordinate system is to place y-axes on both the left and the right sides of a chart, and x-axes above and below. The left and bottom axes usually have numeric or date labels, and the top and right axes just tick marks. This does not have to be the case though; every axis can have axis labels if you want. In general, the axis position in the chart is determined by its intercept. The default value of the intercept is set to the minimums of the coordinate system that the axis is placed in. Adjusting the intercept using the setAxisIntercept method changes the position of the axis in the chart. The axis intercept value is set using units of the coordinate system at right angles to the axis. The example below, extracted from the simplelineplots.LineFill example, places y-axes on both the left and right of the chart.

pTransform1 = new TimeCoordinates();

pTransform1.autoScale(DatasetArray, ChartConstants.AUTOAXES_FAR,

ChartConstants.AUTOAXES_FAR);

pTransform1.setGraphBorderDiagonal(0.125, .08, .92, 0.85) ;

TimeAxis xAxis = new TimeAxis(pTransform1);

xAxis.setColor(Color.white);

gWG.addChartObject(xAxis);

LinearAxis yAxis = new LinearAxis(pTransform1, ChartConstants.Y_AXIS);

yAxis.setColor(Color.white);

gWG.addChartObject(yAxis);

LinearAxis yAxis2 = new LinearAxis(pTransform1, ChartConstants.Y_AXIS);

yAxis2.setAxisIntercept(xAxis.getAxisMax());

yAxis2.setAxisTickDir(ChartConstants.AXIS_MAX);

yAxis2.setColor(Color.white);

gWG.addChartObject(yAxis2);

TimeAxisLabels xAxisLab = new TimeAxisLabels(xAxis );

xAxisLab.setAxisLabelsFormat(ChartConstants.TIMEDATEFORMAT_Y2000);

xAxisLab.setColor(Color.white);

gWG.addChartObject(xAxisLab);

NumericAxisLabels yAxisLab = new NumericAxisLabels(yAxis);

yAxisLab.setColor(Color.white);

yAxisLab.setAxisLabelsFormat(ChartConstants.CURRENCYFORMAT);

gWG.addChartObject(yAxisLab);

The other common reason to have multiple axes in a chart is to delineate the simultaneous use of different coordinate systems in the chart. In this case each coordinate system has an x- and/or y-axis to differentiate it from the other coordinate systems. When the different coordinate systems are created, they usually overlay the same area of the chart. The default positioning of the axes for each coordinate system will all overlay one another, making the axes unreadable. In the y-axis case you will want to offset additional axes to the left, or to the right of the default axis position, using the setAxisIntecept method. When using the setAxisIntercept method, make sure you specify the position using the units of the coordinate system scale at right angles to the axis. Specify an intercept value outside of the normal scale range to offset the axes so that they do not overlap. The example below, extracted from the multipleaxes.MultiAxes example, creates one x-axis, common to all of the charts because the x-scaling for all of the coordinate systems match, and five y-axes, one for each of the five different coordinate systems.

CartesianCoordinates pTransform1 = new CartesianCoordinates( ChartConstants.LINEAR_SCALE, ChartConstants.LINEAR_SCALE);

CartesianCoordinates pTransform2 = new CartesianCoordinates( ChartConstants.LINEAR_SCALE, ChartConstants.LINEAR_SCALE);

CartesianCoordinates pTransform3 = new CartesianCoordinates( ChartConstants.LINEAR_SCALE, ChartConstants.LINEAR_SCALE);

CartesianCoordinates pTransform4 = new CartesianCoordinates( ChartConstants.LINEAR_SCALE, ChartConstants.LINEAR_SCALE);

CartesianCoordinates pTransform5 = new CartesianCoordinates( ChartConstants.LINEAR_SCALE, ChartConstants.LINEAR_SCALE);

LinearAxis xAxis;

LinearAxis yAxis1;

LinearAxis yAxis2;

LinearAxis yAxis3;

LinearAxis yAxis4;

LinearAxis yAxis5;

. // Initialize datasets, coordinate system ranges

// The x-scale range for pTransform1 to pTransform5 are all the same, 0 – 100

// The y-scale range for pTransform1 to pTransform5 are all different

// The plotting area for each pTransform is indentical, leaving a large open

// to the left for extra axes.

pTransform1.setGraphBorderDiagonal(0.35, .15, .9, 0.75) ;

pTransform2.setGraphBorderDiagonal(0.35, .15, .9, 0.75) ;

pTransform3.setGraphBorderDiagonal(0.35, .15, .9, 0.75) ;

pTransform4.setGraphBorderDiagonal(0.35, .15, .9, 0.75) ;

pTransform5.setGraphBorderDiagonal(0.35, .15, .9, 0.75) ;

ChartAttribute attrib1 = new ChartAttribute (Color.blue, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib2 = new ChartAttribute (Color.Red, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib3 = new ChartAttribute (Color.Green, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib4 = new ChartAttribute (Color.Orange, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib5 = new ChartAttribute (Color.Magenta, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib1 = new ChartAttribute (Color.blue, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib2 = new ChartAttribute (Color.red, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib3 = new ChartAttribute (Color.green, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib4 = new ChartAttribute (Color.orange, 2,ChartConstants.LS_SOLID);

ChartAttribute attrib5 = new ChartAttribute (Color.magenta, 2,ChartConstants.LS_SOLID);

xAxis = new LinearAxis(pTransform1, ChartConstants.X_AXIS);

xAxis.setLineWidth(2);

gWG.addChartObject(xAxis);

yAxis1 = new LinearAxis(pTransform1, ChartConstants.Y_AXIS);

yAxis1.setAxisIntercept(0.0);

yAxis1.setChartObjAttributes(attrib1); // axis color matches line color

gWG.addChartObject(yAxis1);

yAxis2 = new LinearAxis(pTransform2, ChartConstants.Y_AXIS);

yAxis2.setAxisIntercept(-18);

yAxis2.setChartObjAttributes(attrib2); // axis color matches line color

gWG.addChartObject(yAxis2);

yAxis3 = new LinearAxis(pTransform3, ChartConstants.Y_AXIS);

yAxis3.setAxisIntercept(-35);

yAxis3.setChartObjAttributes(attrib3); // axis color matches line color

gWG.addChartObject(yAxis3);

yAxis4 = new LinearAxis(pTransform4, ChartConstants.Y_AXIS);

yAxis4.setAxisIntercept(-52);

yAxis4.setChartObjAttributes(attrib4); // axis color matches line color

gWG.addChartObject(yAxis4);

yAxis5 = new LinearAxis(pTransform5, ChartConstants.Y_AXIS);

yAxis5.setAxisIntercept(xAxis.getAxisMax());

yAxis5.setAxisTickDir(ChartConstants.AXIS_MAX);

yAxis5.setChartObjAttributes(attrib5); // axis color matches line color

gWG.addChartObject(yAxis5);

NumericAxisLabels xAxisLab = new NumericAxisLabels(xAxis);

xAxisLab.setTextFont(theFont);

gWG.addChartObject(xAxisLab);

NumericAxisLabels yAxisLab1 = new NumericAxisLabels(yAxis1);

yAxisLab1.setTextFont(theFont);

yAxisLab1.setAxisLabelsFormat(ChartConstants.BUSINESSFORMAT);

gWG.addChartObject(yAxisLab1);

NumericAxisLabels yAxisLab2 = new NumericAxisLabels(yAxis2);

yAxisLab2.setTextFont(theFont);

gWG.addChartObject(yAxisLab2);

NumericAxisLabels yAxisLab3 = new NumericAxisLabels(yAxis3);

yAxisLab3.setTextFont(theFont);

gWG.addChartObject(yAxisLab3);

NumericAxisLabels yAxisLab4 = new NumericAxisLabels(yAxis4);

yAxisLab4.setTextFont(theFont);

gWG.addChartObject(yAxisLab4);

NumericAxisLabels yAxisLab5 = new NumericAxisLabels(yAxis5);

yAxisLab5.setTextFont(theFont);

gWG.addChartObject(yAxisLab5);

Font axisTitleFont = new Font(“SansSerif”, Font.BOLD,12);

AxisTitle xaxistitle = new AxisTitle( xAxis, axisTitleFont, “Event Partition”);

gWG.addChartObject(xaxistitle);

Grid xgrid = new Grid(xAxis, yAxis1,ChartConstants.X_AXIS,

ChartConstants.GRID_MAJOR);

gWG.addChartObject(xgrid);

SimpleLinePlot thePlot1 = new SimpleLinePlot(pTransform1, Dataset1, attrib1);

gWG.addChartObject(thePlot1);

SimpleLinePlot thePlot2 = new SimpleLinePlot(pTransform2, Dataset2, attrib2);

gWG.addChartObject(thePlot2);

SimpleLinePlot thePlot3 = new SimpleLinePlot(pTransform3, Dataset3, attrib3);

gWG.addChartObject(thePlot3);

SimpleLinePlot thePlot4 = new SimpleLinePlot(pTransform4, Dataset4, attrib4);

gWG.addChartObject(thePlot4);

SimpleLinePlot thePlot5 = new SimpleLinePlot(pTransform5, Dataset5, attrib5);

gWG.addChartObject(thePlot5);

Can I add new axes, text objects, plot objects, and images to a chart after it is already displayed; or must I create a new chart from scratch taking into account the additional objects?

There are two ways to add new objects to a chart. The first way is to create all objects when the chart is initially created, but disable the ones that you do not want to show up when the chart is initially rendered. Enable the objects when you want them to show up. Use the chart objects setChartObjEnable method to enable/disable the object. This is useful if you are creating an animated chart where you want the chart to sequence through a predefined series of steps. The second way you add new chart objects to the ChartView using the ChartView.addChartObject method. In both cases you need to call the ChartView.updateDraw() method after any changes are made.

The example below, extracted from the datacursorsandmarkers.DataCursorChart2.CustomChartDataCursor class, creates a new Marker object and NumericLabel object each time a mouse button clicked.

// create marker object at place it at the nearest point

Marker amarker = new

Marker(getChartObjScale(), MARKER_BOX, nearestPoint.getX(), nearestPoint.getY(), 10.0, PHYS_POS);

chartview.addChartObject(amarker);

rNumericLabelCntr += 1.0;

// Add a numeric label the identifies the marker

pointLabel =

new NumericLabel(getChartObjScale(), textCoordsFont, rNumericLabelCntr,

nearestPoint.getX(), nearestPoint.getY(), PHYS_POS, DECIMALFORMAT, 0);

// Nudge text to the right and up so that it does not write over marker

pointLabel.setTextNudge(5,-5);

chartview.addChartObject(pointLabel);

chartview.updateDraw();

How do you zoom charts that use multiple coordinate systems?

The ChartZoom class will zoom one or more simultaneous coordinate systems. The example program zoomexamples.SuperZoom zooms a chart that has one x-axis and five y-axes. Use the ChartZoom constructor that accepts an array of coordinate system objects.

How do you select a chart object and create a dialog panel that permits editing of that objects properties?

The QCChart2D for Java library does not include predefined dialogs for editing chart object properties. The look, feel and details of such dialogs are application specific and it is up to the application programmer to provide these. The property editor tables common to many packages are designed to be used by developers, not end users.

You can add your own dialogs that edit the characteristics important to your end users. If you want to select the chart object by pressing a mouse button while the cursor is on the object, use the FindObj class. Override the mouseClicked method and invoke the appropriate dialog panel there. The following example is extracted from the editchart.EditChart example program.

class CustomFindObj extends FindObj {

CustomFindObj (ChartView component)

{

super( component);

}

public void mouseClicked (MouseEvent event)

{

super.mouseClicked(event);

GraphObj selectedObj = getSelectedObject();

AxisDialog axisdialog = null;

// Call the appropriate chart object dialog panel

if (selectedObj == gWG.xAxis)

axisdialog = new AxisDialog(null, true, xAxis);

else if (selectedObj == gWG.yAxis)

axisdialog = new AxisDialog(null, true, yAxis);

if (axisdialog != null)

{

axisdialog.setLocationRelativeTo(gWG);

axisdialog.setVisible(true);

gWG.updateDraw();

}

CustomFindObj editObject = new CustomFindObj(gWG);

editObject.addFindObjListener();

The AxisDialog classe is derived from the Java JDialog and need to be written by the programmer.. The sample AxisDialog class uses the javax.swing.JColorChooser() to select the color of axis objects

How do you handle missing data points in a chart?

There are two ways to handle missing, or bad data. The first is to mark the data point in the dataset invalid, using the datasets setValidData method. The second is to set the x- and/or y- value of the bad data point to the designated bad data value,ChartConstants.rBadDataValue. Currently this value is set equal to the value of Double.MAX_VALUE. Either method will prevent the data point from being displayed in a chart. If the bad data value is part of a line plot, a gap will appear in the line plot at that point. Bad data points are not deleted from a dataset.

How do you update a chart in real-time?

In general, real-time updates involve adding new objects to a chart, or modifying existing objects that are already in the chart. Once the object is added or changed, call the ChartView.updateDraw() method to force the chart to update using the new values. Objects can be added or modified based on some external event, or in response to a timer event created using java.awt.event.Timer. Make all changes for a given event and call the ChartView.updateDraw method once. The position of most GraphObj derived objects is set or modified using one of the objects setLocation methods. New data points can be added to an existing dataset using one of the datasets addDataPoint, addTimeDataPoint, addGroupDataPoints or addTimeGroupDataPoints methods. ChartPlot derived objects that use datasets will update to reflect the new values when the ChartView.updateDraw method is called. If the coordinates of the new data points are outside of the x- and y-limits of the current coordinate system it may be necessary to rescale the coordinate system so that the new points show up; otherwise the new data points will be clipped. The new scale values can be set explicitly, or calculated using one of the auto-scale methods. The example programs SpectrumAnalyzer, DataLogger, DynPieChart and ScrollingMixedPlot all demonstrate various ways to update charts in real-time.

If you want to change points in an existing dataset, but not the size of the dataset, call the datasets appropriate setXDataValue, setYDataValue, or setDataPoint methods. The dataset has its own copy of the data so you must change these values, not the original values you used to initialize the dataset. If you plan to change every value in the dataset, you can do that point by point, or create a new dataset and swap that in for the old dataset using the plot objects setDataset or setGroupDataset method. Call the ChartView.updateDraw method to force the chart to update using the new values.

How do I prevent flicker when updating my charts on real-time?

Flicker is the result of erasing and redrawing all or part of a chart in the current display buffer. Double buffering of screen images can minimize any flicker. The ChartView class does the actual work of rendering a chart image to the underlying JPanel display buffer. The JPanel class uses double buffering for the display of all screen images. When a chart is updated it is automatically rendered to an off-screen bitmap. When drawing is complete the off-screen bitmap is copied to the screen display buffer, minimizing the effect of flicker.

How do you implement drill down, or data tool tips in a chart?

Implementing drill down or tool tips consists of three major parts:

Trapping a mouse event and determining the mouse cursor position in device and physical coordinates.
Identifying the chart object that intersects the mouse event.
Displaying appropriate information about the chart object.

There are many classes that aid in one or more of these functions. The ChartMouseListener class will trap a mouse event in the chart view. The FindObj class will filter and return the chart object, if any, that intersects the mouse cursor when a mouse button is pressed. The MoveObj class will filter, select and move a chart object as the mouse is dragged across the chart. The DataToolTip class will find the data point in a chart nearest the mouse cursor and display xy information about the data point as a popup ChartText display. The DataToolTip can also be customized for the display of custom information about the selected data point. It only takes a few lines to add a simple y-value tool tip to an existing chart.

DataToolTip datatooltip = new DataToolTip(chartVu);

datatooltip.setEnable(true);

chartVu.setCurrentMouseListener(datatooltip);

Some of the example programs that include tool tips include simplelineplots.LineFill, multiline.Multiline, simplelineplots.LinePlotSegments, multiline.StackedLineChart, logplots.FinLogPlot, bargraphs.SimpleBarChart, bargraphs.GroupBarPlotChart, bargraphs.DoubleBarPlot, financialexamples.OpeningScreen, financialexamples.OHLCFinPlot and piechart.LabeledPieChart.

I do not want to my graph to auto-scale. How do I setup the graph axes for a specific range?

Auto-scaling has two parts. The first is the auto-scaling of the coordinate system based on one or more datasets. The second part is the auto-scaling of the axes that reside in the coordinate system. Manually scale the coordinate system and axes by calling the appropriate constructors. For example:

GregorianCalendar xMin = new GregorianCalendar(1996, ChartConstants.FEBRUARY, 5);

GregorianCalendar xMax = new GregorianCalendar(2002, ChartConstants.JANUARY, 5);

double yMin = 0;

double yMax = 105;

TimeCoordinates simpleTimeScale;

simpleTimeScale = new TimeCoordinates(xMin, yMin, xMax, yMax);

// Create the time axis (x-axis is assumed)

TimeAxis xAxis = new TimeAxis(simpleTimeScale);

// Create the linear y-axis

LinearAxis yAxis = new LinearAxis(simpleTimeScale, ChartConstants.Y_AXIS);

// Create the ChartView object to place graph objects in.

ChartView chartVu = new ChartView();

// Add the x- and y-axes to the chartVu object

chartVu.addChartObject(xAxis);

chartVu.addChartObject(yAxis);

The documentation for the various coordinate system and axis classes includes examples of manual scaling.

How do I update my data, and auto-rescale the chart scales and axes to reflect the new data, after it has already been drawn?

Updating data was discussed in FAQ # 6. If you want the chart to rescale based on the new data, call the appropriate coordinate systems auto-scale method, followed by the auto-axis methods of all related axes. Then call the ChartView.updateDraw method. For example:

// Create the ChartView object to place graph objects in.

TimeSimpleDataset Dataset1 = new TimeSimpleDataset(“Sales”,x1,y1);

TimeCoordinates simpleTimeCoordinates = new TimeCoordinates();

simpleTimeCoordinates.autoScale(Dataset1, ChartConstants.AUTOAXES_FAR , ChartConstants.AUTOAXES_FAR);

ChartView chartVu = new ChartView();

// Create the time axis (x-axis is assumed)

TimeAxis xAxis = new TimeAxis(simpleTimeCoordinates);

// Create the linear y-axis

LinearAxis yAxis = new LinearAxis( simpleTimeCoordinates, ChartConstants.Y_AXIS);

// The following code would be in the code handling the rescale event

// Rescale chart based on a modified Dataset1 datset

simpleTimeCoordinates.autoScale(Dataset1, ChartConstants.AUTOAXES_FAR , ChartConstants.AUTOAXES_FAR);

xAxis.calcAutoAxis();

yAxis.calcAutoAxis();

// Redraw the chart using the rescaled coordinate system and axes

chartVu.updateDraw();

When I use the auto-scale and auto-axis routines my semi-log chart has the log axis scaled using powers of 10 (1, 10,100, 1000, etc.) as the starting and ending values, or as the major tick interval for labeling. How do I make my log graphs start at 20 and end at 50,000, with major tick marks at 20, 200, 2000 and 20000?

The auto-scale routines for logarithmic coordinate systems will always select a power of 10 for the minimum and maximum value of the scale. You can use the auto-scale routine and then override the minimum and/or maximum values for the logarithmic scale. The default LogAxis constructor will pick up on the minimum of the coordinate system and use that as the axis tick mark origin. Or you can leave the coordinate system unchanged, and change the starting point of the axis tick marks using the axis setAxisTickOrigin method. The example below is derived from the logplots.SpeakerSPL example code.

CartesianCoordinates pTransform1 = new CartesianCoordinates( ChartConstants.LOG_SCALE, ChartConstants.LINEAR_SCALE);

pTransform1.autoScale(Dataset1, ChartConstants.AUTOAXES_FAR, ChartConstants.AUTOAXES_FAR);

pTransform1.setScaleStartX(20); // Force start of scale at 20, AutoScale will

// always choose a power of 10 decade.

Background background = new Background( pTransform1, ChartConstants.GRAPH_BACKGROUND, new Color(255,255,255));

gWG.addChartObject(background);

pTransform1.setGraphBorderDiagonal(0.15, .15, .90, 0.8) ;

LogAxis xAxis = new LogAxis(pTransform1, ChartConstants.X_AXIS);

xAxis.setAxisTickOrigin(20);

gWG.addChartObject(xAxis);

How do I create and use custom, multi-line string labels as the axis labels for my graph?

The StringAxisLabels class should be used to create multi-line axis labels. Insert the “\n” new line character to add additional lines to each string used to define the string axis labels. The example below is from the chartaxes.AxisLabelsChart example program.

String xstringlabels[] = {“”,

“Western”+”\n”+”Sales”+”\n”+”Region”,

“Eastern”+”\n”+”Sales”+”\n”+”Region”,

“Southern”+”\n”+”Sales”+”\n”+”Region”,

“Northern”+”\n”+”Sales”+”\n”+”Region”};

StringAxisLabels xAxisLab5 = new StringAxisLabels(xAxis5);

xAxisLab5.setAxisLabelsStrings(xstringlabels,5);

xAxisLab5.setTextFont(graph5Font);

gWG.addChartObject(xAxisLab5);

How do I place more than one graph in a view?

One way to create multiple charts is to create multiple instances of the ChartView class and add each ChartView object to a container object such as a JPanel. A layout manager manages the position and size of each ChartView. Another way is to place multiple charts in the same ChartView object. This makes it easier to guarantee alignment between the axes of separate graphs. The trick to doing this is to create separate coordinate system objects (CartesianCoordinates, TimeCoordinates or PolarCoordinates) for each chart, and to position the plot area of each coordinate system so that they do not overlap. Use one of the coordinate systems setGraphBorder… methods. Many of the examples use this technique, including bargraphs.GroupBarPlotChart, bargraphs.DoubleBarPlot, financialexamples.OHLFinPlot, financialexamples.FinOptions, piecharts.DynPieChart, piecharts.PieAndLineChart and piecharts.PieAndBarChart. The example below was extracted from the financialexamples.OHLCFinPlot class.

pTransform1 = new TimeCoordinates();

pTransform1.setWeekType(weekmode);

pTransform1.autoScale(Dataset1, ChartConstants.AUTOAXES_NEAR, ChartConstants.AUTOAXES_NEAR);

pTransform1.setGraphBorderDiagonal(0.1, .15, .90, 0.6);

pTransform2.setGraphBorderDiagonal(0.1, .7, .90, 0.875) ;

pTransform2 = new TimeCoordinates();

pTransform2.setWeekType(weekmode);

pTransform2.autoScale(Dataset2, ChartConstants.AUTOAXES_NEAR, ChartConstants.AUTOAXES_NEAR);

pTransform2.setGraphBorderDiagonal(0.1, .7, .90, 0.875);

How do I use your software to generate GIF files?

Unlike the JPEG image file format, the GIF file format uses a proprietary data compression algorithm known as LZW. The patent on the LZW compression algorithm is owned by the large computer/data processing company Unisys. Programmers who write commercial applications that use this file format may be subject to paying Unisys royalties. For this reason we do not intent to write any code that explicitly supports the GIF file format. You may be able to seach out some other third party software that handles JPEG to GIF conversions.

Sometimes the major tick marks of an axis are missing the associated tick mark label ?

The axis labeling routines are quite intelligent. Before the label is drawn at its calculated position, the software does a check to see if the bounding box of the new axis label intersects the bounding box of the previous axis label. If the new label is going to overlap the previous label, the label is skipped. You can override this default behavior by calling the objects setOverlapLabelMode method.

setOverlapLabelMode (ChartConstants.OVERLAP_LABEL_DRAW);

Another option, for horizontal axes only, is to stagger the tick mark labels. A stagger automatically alternates the line on which the tick mark label is placed.

setOverlapLabelMode (ChartConstants.OVERLAP_LABEL_STAGGER);

How do I change the order the chart objects are drawn? For example, I want one of my grid objects to be drawn under the charts line plot objects, and another grid object to be drawn top of the charts line plot objects.

There are two ordering methods used to render chart objects. The first method renders the objects in order, as added to the ChartViewobject. Objects added to the view last are drawn on top of objects added first. The second method renders the objects according to their z-order. Objects with the lowest z-order values are rendered first. Objects with equal z-order values are rendered in the ordered they are added to the ChartView object. The second method (z-order rendering) is the default method of object rendering used by the ChartView class. This default behavior can be changed by call the ChartView.setZOrderSortEnable(false) method.

You can change the default z-order value on an object-by-object basis. Call the GraphObj.zetZOrder method to change the z-order for any given object.

See the section in the manual titled Rendering Order of GraphObj Objects for information about the default z-values for all chart objects

The example below sets the z-order value of grid1 to something less than the default value (50) of ChartPlot objects, and the z-order value of grid2 to something greater than the default value.

ChartView chartVu = new ChartView();

Grid grid1 = new Grid(xAxis, yAxis, ChartConstants.Y_AXIS, ChartConstants.GRID_MAJOR);

grid1.setZOrder(40); // This is actually the default value for the grid z-order

chartVu.addChartObject(grid1);

Grid grid2 = new Grid(xAxis, yAxis, ChartConstants.Y_AXIS, ChartConstants.GRID_MINOR);

grid2.setZOder(150); // Grid is drawn after ChartPlot objects

// which have default z-value of 50

chartVu.addChartObject(grid2);

How to I use a ScrollBar object to control horizontal scrolling of the data in my chart?

Since the ChartView class is derived from JPanel, you can place the scroll bar in the ChartView object, or you can place the ChartViewobject and the scroll bar in a parent container and use a layout manager to position everything. If you place the scroll bar in the ChartView you can still position it using a layout manager. The ChartView will always use the entire content area of the underlying JPanel for its canvas and the scroll bars will sit on top of this, not side by side. The example program swingexamples.LinePlotScrollBar uses two scroll bars, a horizontal scroll bar to control scrolling of the x-axis, and a vertical scroll bar that controls the magnitude of the y-axis. You need to implement the AdjustmentListener interface for the derived ChartView in order to process the adjustmentValueChanged method invoked when the scrollbars are moved.

public void updateXScaleAndAxes(int index)

{

int startindex = 0;

startindex = jScrollBar1.getValue();

pTransform1.setScaleStartX( (double) startindex);

pTransform1.setScaleStopX( (double) (startindex + 100));

xAxis.calcAutoAxis();

yAxis.calcAutoAxis();

xAxisLab.calcAutoAxisLabels();

yAxisLab.calcAutoAxisLabels();

this.repaint();

}

public void updateYScaleAndAxes(int index)

{

int startindex = 0;

startindex = jScrollBar2.getValue();

pTransform1.setScaleStartY( (double) -startindex);

pTransform1.setScaleStopY( (double) startindex);

xAxis.calcAutoAxis();

yAxis.calcAutoAxis();

xAxisLab.calcAutoAxisLabels();

yAxisLab.calcAutoAxisLabels();

this.repaint();

}

/** The dummy mouseReleased event method for this object.

* @param event The mouse event contains status information about the mouse.

public void adjustmentValueChanged(AdjustmentEvent e)

{ int startindex = 0;

Object eventobject = e.getSource();

if (eventobject == (Object) jScrollBar1)

{

if (jScrollBar1.getValueIsAdjusting())

return;

startindex = jScrollBar1.getValue();

updateXScaleAndAxes(startindex);

} else if (eventobject == (Object) jScrollBar2)

{

if (jScrollBar2.getValueIsAdjusting())

return;

startindex = jScrollBar2.getValue();

updateYScaleAndAxes(startindex);

}

There are many other examples of Form components interacting with charts. The ContourLinePlot example program uses a CheckBoxobject to specify which contours are to be displayed. The CandlestickWithTable example continuously updates a DataGrid (the base class for the MyDataGrid class) table with open-high-low-close data and simultaneously updates a chart with the same data. The scrolling action of the DataGrid is synchronized with the chart. In the OHLCFinPlot example a Scrollbar controls the time axis of a stock market OHLC chart. The MultiAxes example uses Button objects to select the x-axis range.

I am trying to plot 100,000,000 data points and it takes too long to draw the graph. What is wrong with the software and what can I do to make it faster?

The software runs as fast as we can make it. We do not have any hidden switches that will speed up the software. What you need to do is to step back and think about the best way to display your data.

A fundamental issue that many programmers fail to consider is the relationship between the resolution of the rasterized screen image of the plot and the resolution of the data. A typical chart image will have 500-1000 pixels as the horizontal resolution of the plotting area. This would imply that in the 100M data point example above, every horizontal pixel would represent 50K to 100K data points. Obviously this is a terrible mismatch. In fact it is a bad match for datasets that have more than a couple of thousands points.

So what you do is compress the data before it is displayed. Take the 100M data points and compress them down to 2K data points. The data compression can take several forms. You can take an average of every N points. The resulting dataset will be reduced by a factor of N. You can also find the sum for every N points, the minimum value of every N points, the maximum of every N points, or both the minimum and maximum of every 2N points. The last compression method, minimum and maximum, will always capture any minimums and maximum in the data. The result is that a 2000 point compressed dataset, where there are at least two data points per pixel of horizontal resolution, will look just like the 100,000,000 point dataset, only display hundreds of times faster. The Datset classes all include compression methods (SimpleDataset.compressSimpleDataset, GroupDataset.compressGroupDataset, TimeSimpleDataset.compressTimeSimpleDataset and TimeGroupDataset.compressTimeGroupDataset, TimeGroupDataset.compressTimeFieldSimpleDataset, TimeGroupDataset.compressTimeFieldGroupDataset) that operate on the existing dataset and return a new, compressed dataset. The compressTimeFieldSimpleData and compressTimeFieldGroupDatasetare particular useful because they do not use a fixed sample size of N, instead they compress data so that adjacent time values are an increment of a specific time field (ChartConstants.DAY_OF_YEAR, ChartConstants.WEEK_OF_YEAR, ChartConstants.MONTH, ChartConstants.Year). Compressing data by month and year obviously requires a varying sample size.

Once created, connect the compressed dataset to the ChartPlot object used to display the dataset.

nNumPnts = 1000000;

TimeSimpleDataset RawDataset = new TimeSimpleDataset(“Raw”, xtimedata, ydata,nNumPnts);

int compressXmode = ChartConstants.DATACOMRESS_AVERAGE;

int compressYmode = ChartConstants.DATACOMRESS_MINMAX;

int compressTimeField = Calendar.MONTH;

TimeSimpleDataset CompressedDataset =

RawDataset.compressTimeFileSimpleData( compressXmode,

compressYmode,

compressTimeField,

0, nNumPnts,”Compressed”);

How do I get data from my database into a chart?

The real question is: How do you get data from your database into a simple .Net program, storing sequential data values in data array variables. This is up to you and is independent of the charting software. We recommend that you use the SQL database classes that are part of Java and study the documentation provide by Sun and other sources, such as the O’Reilly programming books. Once you can read individual data elements of your data base it is a trivial matter to place the numeric and calendar data into simple .Net array variables and from there plot the data.

How do I use this charting software to generate chart images “on-the-fly” from a server?

The ChartBufferedImage class creates chart images independent of any physical display context.

Can QCChart2D for Java be used to create programs and controls that run in web browsers?

Yes, we have several examples of how to do this. See Chapter 22 of this manual