xline - draw a partial line

Is there an easy way to get an xline plot for an array of X values, but with the line only taking up the first or last 10% of the Y axis? It's for this plot, where plotting the whole line is obscuring the blue data too much..

This is for a script where the Y axis is logarithmic and its values are not always the same. I could plot the lines in a loop, having calculated the appropriate Y values. But is there a simpler way?

[dormant]

Thanks everyone, I'll have to bite the bullet and do it manually.

[Voss]

@dormant: How about putting the black vertical lines beneath the blue line, so that the blue line is not obscured? That can be done: (1) by plotting the black lines before the blue line, or (2) by reordering the lines in the axes Children property after they are plotted, e.g. using uistack or setting the Children property directly.

Example:

% plot a thick line and a thin line on top:
figure()
h_thick = plot(1:10,'LineWidth',10);
hold on
h_thin = plot(1:10,'LineWidth',2);

% thin line is 1st in axes Children (most recently created):
ax = gca();
ch = get(ax,'Children');
ch == h_thin

ans = 2×1 logical array
   1
   0

% new figure and axes (only necessary for demonstration in Answers):
ax = copyobj(ax,figure());

% get the new handle of the thin line (also only necessary for
% demonstration in Answers):
ch = get(ax,'Children');
h_thin = ch(1);

% reverse the order of the axes Children:
set(ax,'Children',flip(get(ax,'Children')));

% now the thin line is underneath the thick one,
% and the thin line is last in Children:
ch = get(ax,'Children');
ch == h_thin

ans = 2×1 logical array
   0
   1

[dormant]

That doesn't work with my data. But thanks for the suggestion.

Answer 1

[Dyuman Joshi]

You will have to do that manually -

%Sinosuidal data for example
x = 0:0.01:10;
y = sin(x);

plot(x, y, 'LineWidth', 2)

%Generate random values to put xlines
X = randi(20, 1, 10)/2

X = 1×10

10.0000 5.0000 7.5000 8.5000 2.5000 5.0000 1.5000 8.5000 9.5000 4.0000

y1 = -0.2;
y2 = 0.3;

%Draw xlines for the values in X, extending from y1 upto y2
hold on
plot([X; X], [y1; y2], 'k')

[Voss]

@dormant: Note that this method produces one line object for each x value. As you can see below h is an array of 10 lines:

figure; rng(11)

% Sinosuidal data for example
x = 0:0.01:10;
y = sin(x);

plot(x, y, 'LineWidth', 2)

% Generate random values to put xlines
X = randi(20, 1, 10)/2

X = 1×10

2.0000 0.5000 5.0000 7.5000 4.5000 5.0000 0.5000 5.0000 9.5000 9.0000

y1 = -0.2;
y2 = 0.3;

% Draw xlines for the values in X, extending from y1 up to y2
hold on
h = plot([X; X], [y1; y2], 'k') % 10 lines

h = 
  10×1 Line array:

  Line
  Line
  Line
  Line
  Line
  Line
  Line
  Line
  Line
  Line

Having too many graphics objects that MATLAB needs to render can make your GUI sluggish to interact with. Therefore, in general it's a good idea to minimize the number of graphics objects you create.

To that end, you can achieve the same effect as above with a single line object by taking advantage of how NaNs are represented in lines (NaNs make gaps). Here's the same example, but with one line object for all vertical lines:

figure; rng(11)

% Sinosuidal data for example
x = 0:0.01:10;
y = sin(x);

plot(x, y, 'LineWidth', 2)

% Generate random values to put xlines
X = randi(20, 1, 10)/2

X = 1×10

2.0000 0.5000 5.0000 7.5000 4.5000 5.0000 0.5000 5.0000 9.5000 9.0000

y1 = -0.2;
y2 = 0.3;

% Draw xlines for the values in X, extending from y1 up to y2
hold on

N = numel(X);
X_plot = [X; X; NaN(1,N)];
Y_plot = repmat([y1; y2; NaN],1,N);
h = plot(X_plot(:),Y_plot(:), 'k') % one line

h = 
  Line with properties:

              Color: [0 0 0]
          LineStyle: '-'
          LineWidth: 0.5000
             Marker: 'none'
         MarkerSize: 6
    MarkerFaceColor: 'none'
              XData: [2 2 NaN 0.5000 0.5000 NaN 5 5 NaN 7.5000 7.5000 NaN 4.5000 4.5000 NaN 5 5 NaN 0.5000 0.5000 NaN 5 5 NaN 9.5000 9.5000 NaN 9 9 NaN]
              YData: [-0.2000 0.3000 NaN -0.2000 0.3000 NaN -0.2000 0.3000 NaN -0.2000 0.3000 NaN -0.2000 0.3000 NaN -0.2000 0.3000 NaN -0.2000 0.3000 NaN -0.2000 0.3000 NaN … ] (1×30 double)

  Use GET to show all properties

Answer 2

[Mathieu NOE]

hello

I am not aware that a special version of xline exist , but you can do your own special xline like that (a very crude and simple code that you can save as a function for future massive use) :

simply try with different a and b values to adjust the amount of blank around the data curve

% dummy data
n = 1000;
t = (0:n-1)/n;
y = 0.1 + sin(pi*t/max(t)).^3.*abs(sin(2*pi*10*(t+1*t.^3))); % signal

[PKS,LOCS] = findpeaks(y,t,'SortStr','descend','NPeaks',15);

figure(1)
ylimits = [1e-2 1e1];
semilogy(t,y,'b',LOCS,PKS,'.r','linewidth',2,'markersize',25);grid on
% add my special vertical lines
a = 0.25; % factor on the lower portion (must be < 1)
b = 1.25; % factor on the upper portion (must be > 1) 
hold on
for k = 1:numel(PKS)
    plot(LOCS(k)*ones(2,1), [ylimits(1) a*interp1(t,y,LOCS(k))],'--k','linewidth',0.5);
    plot(LOCS(k)*ones(2,1), [b*interp1(t,y,LOCS(k)) ylimits(2)],'--k','linewidth',0.5);
end
ylim(ylimits);
hold off

Answer 3

[Image Analyst]

You can use line or plot instead of xline

Answer 4

[dormant]

I decided on an alternative approach. I plotted the data again with another axis with tick marks where my xlines would be.

I'm not sure the axis properties are set optimally, but this does the job.

x = 1:100;
y = sin(x/10);

figure;
t = tiledlayout(1,1);

ax1 = axes(t);
plot(x,y,'k-');

ax2 = axes(t);
plot(ax2,x,y,'k-');
ax2.XAxisLocation = 'top';
xtick( [7 8 9 12 73 84] );
ax2.XTickLabel = {};
ax2.TickDir = 'in';
ax2.XColor = 'r';
ax2.YTick = [];
ax2.XAxisLocation = 'top';
box( ax2, 'XColor', 'k', 'YColor', 'k' );
ax2.TickLength = [0.1 0];

% Link the axes
linkaxes([ax1 ax2], 'x')

[Image Analyst]

Looks confusing to me unless it's explained somewhere. The red lines don't go to the curve but all go to some constant y value. And the x location of them seems random.