MATLAB: Combine Multiple Plot Components into Single Plot Legend Entry

plotplot legend

I am trying to prepare a plot legend for a 2×2 plot figure. Each subplot has multiple plot components. Some these plot components are related; for example, I use "scatter" to plot dataset points and then "plot" to join them with a line of the same color. However, when I try to create a plot legend, I can never combine multiple plot components into a single plot legend entry, so the plot legend shows the a point with a line going through it instead of just the point itself. An example figure I am working with is attached.

A quick example of the code I use is below.

For the top left subplot:

scatter(hour2,ph2,30,'o','filled','MarkerFaceColor',[0 0 0.8]); %Entry 1

hold on

plot(hour2,ph2,'Color',[0 0 0.8],'LineWidth',1,'LineStyle','-'); %Entry 2

hold off

Then:

plotleg = legend('Entry 1','Entry 2,'Location','bestoutside','Orientation','vertical');

legend('boxon');

set(plotleg,'FontSize',11);

In that example, each will appear as its own plot legend entry. Is it possible to modify that example code in any way so that the scatter+plot components appear as a single plot legend entry or is there a better way to accomplish this altogether? Thanks in advance for the help.

Best Answer

Instead of using both 'scatter' and 'plot', why not just use 'plot' to create a line plot with markers?

You'd change the call to 'plot' in your example as follows:

plot(hour2,ph2,'-o','Color',[0 0 0.8],'LineWidth',1,'MarkerFaceColor',[0 0 0.8]);

Then your legend will only have one entry for the plot. It sounds like this is what you are trying to accomplish.

Related Solutions

MATLAB: Modify legend markers in plot

http://www.mathworks.com/matlabcentral/answers/277438-marker-and-line-in-legend-in-matlab-plot#answer_216684

MATLAB: Hiding Plot Elements makes those elements appearing in the legend and the legend itself a much lighter shade than it should be. How to fix this

OK, wasn't too hard to take yours and make an example...

figure(1)
colormap jet
hAx(1) = subplot(5,5,[1,2,6,7]);
yyaxis left
ylim([5.6 26.4]);
set(gca,'FontSize',10);
set(gca,'ycolor','k');
set(gca,'ytick',[8 12 16 20 24]);
ytix = get(gca,'ytick');
ytixlbl = regexp(sprintf('%.0f\n',ytix), '\n', 'split');
set(gca,'yticklabel',ytixlbl(1:end-1));
ylabel('Salinity','FontSize',10,'FontWeight','bold');
text(7.0,26.6,'a','FontSize',18,'FontWeight','bold');
str1 = {'01 June 2016'};
text(14.8,25,str1,'FontSize',10,'FontWeight','bold');
hour1=linspace(9,19,7); sal1=[22 21 15 9 12 22 23];
hAr(1)=area(hour1,sal1,'FaceColor',[0.8300 0.8300 0.8300],'EdgeColor',[0.8300 0.8300 0.8300],'LineStyle','-');
yyaxis right
ph1=[8 10 72 68 8 7 2];
hL(1,1)=plot(hour1,ph1,'o-','MarkerSize',7,'MarkerFaceColor',[0 0 0.8],'Color',[0 0 0.8],'LineWidth',1);
ylim([-10.8 82.8]);
set(gca,'FontSize',10);
set(gca,'ycolor','k');
set(gca,'ytick',[0 18 36 54 72]);
ytix = get(gca,'ytick');
ytixlbl = regexp(sprintf('%.0f\n',ytix), '\n', 'split');
set(gca,'yticklabel',ytixlbl(1:end-1));
ylabel('[H^{+}_{T}] (nmol kg^{-1})','FontSize',10,'FontWeight','bold');
xlim([8.5 19.5]);
set(gca,'xtick',[9,11,13,15,17,19]);
xlabel('Hour','FontSize',10,'FontWeight','bold');
set(gca,'FontSize',10);
box on
hold on
dum1=nan(size(hour1));
hL(2,1)=plot(hour1,dum1,'s-','Color',[0.5500 0.7100 0.0000],'LineWidth',1,'MarkerSize',7,'MarkerFaceColor',[0.5500 0.7100 0.0000]);
hL(3,1)=plot(hour1,dum1,'>-','Color',[0.9100 0.4100 0.1700],'LineWidth',1,'MarkerSize',6,'MarkerFaceColor',[0.9100 0.4100 0.1700]);
hLg=legend([hAr; hL],'Salinity','[H^{+}_{T}]','DIC','TA','Location','bestoutside','Orientation','vertical');
return

The above adds the dummy linestyles from which to create the legend onto the first suplot(). Just use the handles to the line/area objects desired to get the right linestyles.

NB: You don't need scatter, just use the same properties of the line--that reduces clutter by quite a bit.

Above produces:

The size/location will come back when you use figure size than just the default figure window size here on screen.

The key item is the NaN data to put the desired linestyles on the figure for legend. This is a most useful "trick" to know for many purposes with HG.

Related Question