Matlab - how to make a custom legend - matlab

I have the following picture :
And I would like to make a legend for it. Basically, I want to make a legend for each type of rectangle. In the legend box, I want to mark each color line according to the type of body which it marks:
green line : head
yellow line : torso
purple line : right arm
cyan line : left arm
red line : left leg
blue line : right leg
This is basically custom, because I have more rectangles of each type. How can I do a custom legend and attach it to the figure which draws this picture?

There are 2 ways you could go about this. You could create your squares and then assign them to an hggroup. This way you dont have multiple items for each color. Something like this:
hold on
for ii = 1:4
hb(ii) = plot(rand(1,2), rand(1,2),'color','r');
end
hg = hggroup;
set(hb,'Parent',hg)
set(hg,'Displayname','Legs')
legend(hg)
Or you could create dummy objects, like this:
hold on
for ii = 1:4
hb(ii) = plot(rand(1,2), rand(1,2),'color','r');
end
p = plot([],[],'r');
legend(p,'Legs')
The former is a little more elegant.

I would like to add to dvreed77's answer on using hggroup that for clean legend use, I also needed to set the 'IconDisplayStyle' (Matlab R2014a), such that:
%4 kinds of lines:
n_areas = 4;
n_lines = 10;
%use built-in color map
cmap = hsv(n_areas);
%plot lines and generate handle vectors
h_fig = figure;
hold on
h_lines = zeros(1,n_lines);
for l = 1:n_areas
for k = 1:n_lines
h_lines(k) = plot(rand(1,2), rand(1,2),'Color',cmap(l,:));
end
%Create hggroup and set 'icondistplaystyle' to on for legend
curPlotSet = hggroup;
set(h_lines,'Parent',curPlotSet);
set(get(get(curPlotSet,'Annotation'),'LegendInformation'),...
'IconDisplayStyle','on');
end
%Now manually define legend label
legend('heads','legs','hands','feet')

The simplest way I can think of is to first plot one rectangle of each type and construct a legend for only unique rectangles. Like so:
figure;
hold on;
% unique rectangles
plot(rand(1, 10), 'b');
plot(rand(1, 10), 'g');
% the rest
plot(rand(1, 10), 'b');
plot(rand(1, 10), 'g');
% use normal legend with only as many entries as there are unique rectangles
legend('Blue', 'Green');
You will have many lines of the same color, but a legend only for unique colors.

Just draw legend dots outside the plot:
figure;
plot(-1,-1,'gs',-1,-1,'b^',-1,-1,'ro');
legend('x1','x2','x3','Location','NorthWest');
xlim([0,1]); ylim([0,1]);

To control the appearance of legend entries, plot points that have values which are NaN then pass the objects returned by plot and an array of labels to the legend function (NaN points are not visible in the plot, but appear in the legend).
colors = ["red", "blue"];
labels = ["this is red", "this is blue"];
% We 'plot' a invisible dummy point (NaN values are not visible in plots),
% which provides the line and marker appearance for the corresponding legend entry.
p1 = plot(nan, nan, colors(1));
hold on
p2 = plot(nan, nan, colors(2));
% Plot the actual plots. You can change the order of the next two function calls
% without affecting the legend.
plot([0, 1], [0, 1], colors(1));
plot([0, 1], [1, 0], colors(2));
legend([p1, p2], labels)
If the plots in [p1, p2] are not in the current figure when legend([p1, p2], labels) is called, then it will raise the following error:
Invalid argument. Type 'help legend' for more information.
You can filter plots that are not in the current figure using something like this:
plots_in_figure = findall(gcf,'Type','Line');
plots_for_legend_indices = ismember([p1, p2], plots_in_figure);
plots_for_this_legend = this.plots_for_legend(plots_for_legend_indices);
legend(plots_for_this_legend, labels)

Related

Can we add legend for function "text" to matlab?

I want to add legend for my plot. Because I want to use the marker plot 'heartsuit', I use the 'text' function. If I add legend function in my code, it can't work. The command window say that 'Warning: Plot empty.' So, can we add legend to 'text' function? I have searched in many source, and I cannot find it.
clear all;
clc;
m = '\heartsuit';
x = 0:pi/5:2*pi;
y = sin(x);
text(x,y,m,'fontname','Arial','color','red','FontSize',18,'HorizontalAlignment','center','VerticalAlignment','middle');
grid on;
xlim([min(x) max(x)])
ylim([min(y) max(y)])
legend('Solusi Numerik');
Here's a hack. Plot a fake NaN point, create a legend for it, hide its legend line, and add the heart-suit in the string with appropriate space at an appropriate position. Adjust the color of heart-suit and/or string if needed.
hold on;
LgdStr = 'Solusi Numerik'; %Your legend string
hNaN = plot(NaN,NaN); %Plotting nothing
[~, icons] = legend(hNaN, LgdStr);%Creating a legend to get required space for string
icons(2).Visible = 'off'; %Hiding the fake legend line
icons(1).Position(1) = 0.125; %Adjusting the starting position of text
icons(1).String = ['\color{red}', m, ' \color{black}',LgdStr];
%Last line includes red-colored heart-suit at reasonable space from black-colored text
Result:

How to add an independent text in MATLAB plot legend

I need an additional text in the legend that is not related with graphical data together with the legend captions.
Something like this (it was made in OriginLab):
Following to this link Add custom legend without any relation to the graph
I can add some text using plot(NaN,NaN,'display','add info here2', 'linestyle', 'none'). But there is an indent in this text:
How to avoid this indent? And is there a more elegant method to add the text that is not associated with the legend together with the legend captions?
The legend function will return as its second output argument handles for all of the components that make up the symbols and text in the legend. You can therefore plot "dummy" lines as placeholders in the legend, reorder the handles when creating the legend to put the text where you want it, and modify the legend objects accordingly. Here's an example:
x = linspace(0, 2*pi, 100);
hl = plot(x, [sin(x); cos(x); tan(x); nan(size(x))].'); % Add a line of NaNs
axis([0 2*pi -4 4]);
[~, objH] = legend(hl([1 2 4 3]), 'sin', 'cos', 'junk', 'tan'); % Reorder handles
set(findobj(objH, 'Tag', 'junk'), 'Vis', 'off'); % Make "junk" lines invisible
pos = get(objH(3), 'Pos'); % Get text box position
set(objH(3), 'Pos', [0.1 pos(2:3)], 'String', 'also...'); % Stretch box and change text
You can use annotations. It's not perfect, but with few adjustments you'll get what you want. Here's an example:
% somthing to plot:
x = [0:0.1:5; 5:0.1:10].';
y = sin(x);
% plot the real data:
plot(x,y);
hold on
% make some space in the legend:
Spacing_lines = 3;
h = plot(nan(size(x,1),Spacing_lines));
hold off
set(h,{'Color'},{'w'}); % clear the dummy lines
% place the legend:
hl = legend([{'lable1','lable2'} repmat({''},1,Spacing_lines)]);
% add your text:
annotation('textbox',hl.Position,'String',{'Some info';'in 2 lines'},...
'VerticalAlignment','Bottom','Edgecolor','none');
And from this you get:
You can just add any text to any point of plot in this way:
txt1 = 'some information';
text(x1,y1,txt1)
where x1, y1 - coordinates.
By the way function text function has a lot of different properties (colors, font size, alignment etc.).
I think the easiest way is to just create some dummy function, plot it but set the color="none" - that way it will only show up in the legend (if that is where you wanted it).

Matlab: same colormap for hist, plot and mesh

I have datasets for 5 different frequencies in a matrix and I want to illustrate them using plot, hist and mesh. However, each plot type is using different colormaps (see picture), so I need a legend for every plot.
Is there a way to set the same colormap for all plot types, or specify one for each of them? Another thing thats strange: I can set the colormap for hist using the figure tools for example, but not for the normal plot. For the mesh I have to use a hold on loop, so I guess setting the color here is different from defining a colormap?
Edit:
Here is a minimal example. Its still not working, see comments in the code below.
clear all;
close all;
clc;
% make up some data with the original format
freqLen = 5;
data = zeros(10, 3, 3, freqLen);
data(:, :, :, 1) = rand(10, 3, 3);
data(:, :, :, 2) = rand(10, 3, 3)+1;
data(:, :, :, 3) = rand(10, 3, 3)+2;
data(:, :, :, 4) = rand(10, 3, 3)+3;
data(:, :, :, 5) = rand(10, 3, 3)+4;
% reshape data so we get a vector for each frequency
dataF = reshape(data, [10*3*3, freqLen]);
% prepare colors for plot, try to get 5 colors over the range of colormap
% but I get wrong colors using both methods below!
%cols = colormap(jet);
%cols = cols(1:round(length(cols)/length(freqGHz)):end, :);
cols = jet(freqLen);
% plot samples in 3D
figure('Position', [0 0 1000 1000]);
subplot(211);
hold on;
for iF = 1:freqLen
dataThisF = dataF(:, iF);
data3D = reshape(dataThisF, [10*3, 3]);
mesh(data3D);
% try to give each "holded" mesh a different color. Not working!
% after the loop, all meshes have the last color
set(get(gca, 'child'), 'FaceColor', 'w', 'EdgeColor', cols(iF, :));
end
view(60, 20);
% plot samples
subplot(223);
hold on;
for iF = 1:freqLen
% the loop is not avoidable
% because matlab maps the colors wrong when plotting as a matrix
% at least its not using the colormap colors
plot(dataF(:, iF), 'Color', cols(iF, :));
end
% plot histogram
subplot(224);
% actually the only one which is working as intended, horray!
hist(dataF, 50);
How can I give a holded mesh a single color, different from the others? How can I map the correct jet colormap when plotting a matrix using simple line plot, or at least get 5 colors from the jet colormap (jet(5) give 5 different colors, but not from start to end)?
What you are talking about is mainly the ColorOrder property (and not the colormap of the figure).
The link for the colororder given just above will explain you how to force Matlab to use a given set of colors for all the plots. It works perfectly fine for plot. You wouldn't need a loop, just define the DefaultColororder property of the figure before the plots then plot all your series in one single call, Matlab will assign the color of each plot according to the order you defined earlier.
For mesh and hist it is not that simple unfortunately, so you will have to run a loop to specify the color or each graphic object. To modify a graphic object property (like the color) after it has been created, you have to use the set method, or even the direct dot notation if you're using a Matlab version >= 2014b. For both methods you needs to have the handle of the graphic object, so usually the easiest when you know you'll need that is to retrieve the graphic object handle at the time of creation*.
*instead of dirty hack like get(gca, 'child'). This is quite prone to errors and as a matter of fact was wrong in your case. Your code wouldn't color properly because you were not getting the right graphic handle this way.
The code below plots all your graphs, retrieve the handle of each graphic object, then assign the colors in the final loop.
%// Get a few colors
cols = jet(freqLen);
% plot samples in 3D
figure('Position', [0 0 1000 1000]);
set( gcf , 'DefaultAxesColorOrder',cols) %// set the line color order for this figure
subplot(2,1,1,'NextPlot','add'); %// 'NextPlot','add' == "hold on" ;
for iF = 1:freqLen
dataThisF = dataF(:, iF);
data3D = reshape(dataThisF, [10*3, 3]);
h.mesh(iF) = mesh(data3D) ; %// plot MESH and retrieve handles
%// You can set the color here direct, or in the last final "coloring" loop
%// set( h.mesh(iF) , 'FaceColor', 'w', 'EdgeColor', cols(iF, :));
end
view(60, 20);
%// plot samples
subplot(223);
h.plots = plot(dataF); %// plot LINES and retrieve handles
%// plot histogram
subplot(224);
[counts,centers] = hist(dataF, 50 ) ; %// get the counts values for each series
h.hist = bar(centers,counts) ; %// plot HISTOGRAM and retrieve handles
%// now color every series with the same color
for iF = 1:freqLen
thisColor = cols(iF, :) ;
set( h.mesh(iF) , 'EdgeColor' , thisColor , 'FaceColor', 'w' );
set( h.hist(iF) , 'EdgeColor' , thisColor , 'FaceColor' , thisColor )
%// this is actually redundant, the colors of the plots were already right from the
%// beginning thanks to the "DefaultColorOrder" property we specified earlier
set( h.plots(iF) , 'Color' , thisColor )
end
Will land you the following figure:
UPDATE: The original question asked
Is there a way to set the same colormap for all plot types, or specify one for each of them?
this answer, answers that question with colormap taken to mean colormap in MATLAB literally.
You can set the colormap for a figure using colormap. You could use one of the many built in colormaps or specify your own.
An example using mesh and the builtin hsv colormap could be
figure;
mesh(data);
colormap(hsv);
This would apply a colormap based on
to your figure. You can also create your own colormap like
map = [1, 0, 0,
1, 1, 1,
0, 1, 0,
0, 0, 0];
colormap(map);
which would create a colormap with the colours Red, White, Blue and Black.
The MATLAB documentation contains extensive information on the use of colormap.

Xtick marks and Xtick labels on heatmap in Matlab

Environment: Windows 7 64 bit, Matlab 2014a
Objective:
To draw a heatmap of errors for two parameters to be optimized.
To put proper tick marks and tick values
Draw the grid lines in the correct place
Problem: Arranging the X and Y tick positions and values. When the last ("end") value of the vectors in x and y axes are the same, the code I use puts the ticks and values properly. However, when the end values are different it does not, and produces something really weird.
Below I have included the code which I have modified so that you can run it without the need of adding anything. Of course in my case the error vector are the error values, not random numbers. To see the problem of "end value" use the second b vector.
fontsize = 20
k = [2^-5, 2^-3, 2^-1, 2^0, 2^1, 2^3, 2^5, 2^7, 2^9, 2^11, 2^13, 2^15]
b = [2^-5, 2^-3, 2^-1, 2^0, 2^1, 2^3, 2^5, 2^7, 2^8, 2^9, 2^10, 2^11, 2^13, 2^15]
% b = [2^-5, 2^-3, 2^-1, 2^0, 2^1, 2^3, 2^5, 2^7, 2^8, 2^9, 2^10, 2^11, 2^13, 2^19]
errorVector = randi(20, 1, length(b)*length(k))'
figure
% Create a matrix from error vector (size of error vector is [length(k)*length(b),1])
B = reshape(errorVector, [length(b), length(k)])
B = flipud(B)
% imagesc(x,y,C)
imagesc(b, k, B)
title('Heatmap Parameters Percent Error', 'FontSize', fontsize);
% Set colorbar limits
caxis([0 15])
colorbar;
ax1 = gca;
xTickLabel = (k)'
xTick = linspace(k(1), k(end), numel(xTickLabel))';
set(ax1, 'XTick', xTick, 'XTickLabel', xTickLabel)
xlabel('k parameter', 'FontSize', fontsize)
yTickLabel = (b)'
yTick = linspace(b(1), b(end), numel(yTickLabel))';
set(ax1, 'YTick', yTick, 'YTickLabel', flipud(yTickLabel(:)))
ylabel('b parameter', 'FontSize', fontsize)
set(ax1,'FontSize', fontsize)
Here, change any of the end values of b or k vectors, and the program will output a graph where the X and Y ticks are totally wrong.
Also, I would like to draw grid lines. When I use "grid on" it draws grid lines right on the tick marks which is not correct in the case of heatmap. Because tick marks will be in the center of the columns -or rows- but the grid lines should be at the boundaries between the columns -or rows.
By the way if you know a better way to plot a heatmap in Matlab, please do tell.
Please help me solve this problem. Any help is appreciated,
Ilyas
First of all - you don't need to flipud the B matrix - by default imagesc plots the y-data inversed, but you can fix this with the following statement:
set(gca,'ydir','normal');
This also means you don't have to mess around with flipping the tick-labels. You can get the labels right by doing the following:
% replace the imagesc call with:
imagesc(B);
set(gca,'ydir','normal');
% formatting stuff
...
% replace the set commands with:
set(ax1, 'XTick', 1:length(k), 'XTickLabel', k)
set(ax1, 'YTick', 1:length(b), 'YTickLabel', b)
By default, if you don't provide x and y data to the imagesc command, it will number them linearly (1,2,3...). Basically what we've done here is make sure that it has ticks for each of the elements of b and k, and then set the labels to the values of the respective vectors.
Unfortunately, I'm not sure if there is a way to get the grid spacing right with imagesc or not. You could try using pcolor instead, which has it's own set of issues, but allows you to get grid lines (of sorts) between the elements. It also allows you to use an interpolated shading mode, which will make your plot look more like a typical heat map.
To use pcolor instead, you just have to replace imagesc:
% imagesc(B);
% set(gca,'ydir','normal');
pcolor(B);
% this uses a smoother shading method, for a more 'heatmap' like output
% shading interp
Everything else should work as intended, I believe.

Minor grid with solid lines & grey-color

I'm using the following to display the minor grid in my plot:
grid(gca,'minor')
set(gca,'MinorGridLineStyle','-')
but I'd like to change the color of the grid lines to a nice greyscale. I can't find any option 'grid color' in matlab... Do you know any or any workaround?
I found this: http://www.mathworks.com/matlabcentral/fileexchange/9815-gridcolor but as I read of the comments, it doesn't work very well and further it only changes gridcolor, not the color of the minor grid...
Thanks!
EDIT:
Problem with semilogx as posting here now:
x = [1e-9 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2]';
y1 = linspace(20, 90, 8);
y2 = y1.^2;
y3 = y1./y2+5;
% plotte: http://www.mathworks.com/help/techdoc/ref/linespec.html
myfig = figure('Position', [500 500 445 356]); %[left, bottom, width, height]:
p1 = semilogx(x,y1,'x--r',x,y2,'*-b');
ax1 = gca;
set(ax1, 'Position',[0.13 0.18 0.75 0.75]);
xlim([0 max(x)]);
ylim([0 max([max(y1) max(y2)])]);
col=.85*[1 1 1];
%# create a second transparent axis, same position/extents, same ticks and labels
ax2 = axes('Position',get(ax1,'Position'), ...
'Color','none', 'Box','on', ...
'XTickLabel',get(ax1,'XTickLabel'), 'YTickLabel',get(ax1,'YTickLabel'), ...
'XTick',get(ax1,'XTick'), 'YTick',get(ax1,'YTick'), ...
'XLim',get(ax1,'XLim'), 'YLim',get(ax1,'YLim'),...
'XScale', 'log');
%# show grid-lines of first axis, give them desired color, but hide text labels
set(ax1, 'XColor',col, 'YColor',col, ...
'XMinorGrid','on', 'YMinorGrid','on', ...
'MinorGridLineStyle','-', ...
'XTickLabel',[], 'YTickLabel',[],'XScale', 'log');
%# link the two axes to share the same limits on pan/zoom
linkaxes([ax1 ax2],'xy');
Displaying like this:
EDIT2: A problem occurs when adding a second y-axes as in the following picture, look at the ticks of the right y-axes:
this will be discussed here to have a better overview!
Matlab: Problem with ticks when setting minor grid style and two y-axis
Set the 'XColor','YColor' axes properties. Note that these properties determine the color of the axis lines, tick marks, tick mark labels, and the axis grid lines, so AFAIK you can't assign those different colors than that of the entire axis..
Example:
plot(rand(10,1))
set(gca, 'XMinorGrid','on', 'YMinorGrid','on', 'XColor','r', 'YColor','g')
EDIT1:
You can always create a second transparent axis with the desired grid colors, but with no ticks or labels, stacked on top of the current axis. Here is an example:
%# create plot as usual
plot(rand(10,1))
hAx1 = gca;
%# create a second axis, same position/extents, no tick or labels, colored grid-lines
hAx2 = axes('Position',get(hAx1,'Position'), ...
'Color','none', 'TickLength',[1e-100 1e-100], ...
'XMinorGrid','on', 'YMinorGrid','on', ...
'Box','off', 'XColor','g', 'YColor','r', ...
'XTickLabel',[], 'YTickLabel',[], ...
'XTick',get(hAx1,'XTick'), 'YTick',get(hAx1,'YTick'), ...
'XLim',get(hAx1,'XLim'), 'YLim',get(hAx1,'YLim'));
%# position it on top
%#uistack(hAx2,'top')
%# redraw the enclosing box in the original axis colors
x = get(hAx1,'XLim');
y = get(hAx1,'YLim');
line([x([1 2]) nan x([2 1])],[y([1 1]) nan y([2 2])],'Color',get(hAx1,'XColor'))
line([x([1 1]) nan x([2 2])],[y([1 2]) nan y([2 1])],'Color',get(hAx1,'YColor'))
The only problem is that the grid lines are drawn on top of your plot, which might get in the way if the grid-lines are thick :)
EDIT2:
Seems like #yoda had a similar idea to the above. Here is a slightly improved version inspired by his solution:
%# create plot as usual
plot(11:20, rand(10,1)*5)
hAx1 = gca; %# get a handle to first axis
%# create a second transparent axis, same position/extents, same ticks and labels
hAx2 = axes('Position',get(hAx1,'Position'), ...
'Color','none', 'Box','on', ...
'XTickLabel',get(hAx1,'XTickLabel'), 'YTickLabel',get(hAx1,'YTickLabel'), ...
'XTick',get(hAx1,'XTick'), 'YTick',get(hAx1,'YTick'), ...
'XLim',get(hAx1,'XLim'), 'YLim',get(hAx1,'YLim'));
%# show grid-lines of first axis, give them desired color, but hide text labels
set(hAx1, 'XColor','g', 'YColor','r', ...
'XMinorGrid','on', 'YMinorGrid','on', ...
'XTickLabel',[], 'YTickLabel',[]);
%# link the two axes to share the same limits on pan/zoom
linkaxes([hAx1 hAx2],'xy');
%# lets create a legend, and some titles
legend(hAx1, 'text')
title('title'), xlabel('x'), ylabel('y')
EDIT3 (take 2):
Here is the same example but with a log-scale x-axis. Note how instead of creating a second axis and manually setting its properties to match the first, I simply copyobj the axis, and delete its children.
%# create a plot as usual (x-axis is in the log-scale)
semilogx(logspace(0,5,100), cumsum(rand(100,1)-0.5))
xlabel('x'), ylabel('y'), title('text')
legend('plot')
%# capture handle to current figure and axis
hFig = gcf;
hAx1 = gca;
%# create a second transparent axis, as a copy of the first
hAx2 = copyobj(hAx1,hFig);
delete( get(hAx2,'Children') )
set(hAx2, 'Color','none', 'Box','on', ...
'XGrid','off', 'YGrid','off')
%# show grid-lines of first axis, style them as desired,
%# but hide its tick marks and axis labels
set(hAx1, 'XColor',[0.9 0.9 0.9], 'YColor',[0.9 0.9 0.9], ...
'XMinorGrid','on', 'YMinorGrid','on', 'MinorGridLineStyle','-', ...
'XTickLabel',[], 'YTickLabel',[]);
xlabel(hAx1, ''), ylabel(hAx1, ''), title(hAx1, '')
%# link the two axes to share the same limits on pan/zoom
linkaxes([hAx1 hAx2], 'xy');
%# Note that `gca==hAx1` from this point on...
%# If you want to change the axis labels, explicitly use hAx2 as parameter.
You should get the correct plot in your example with this code. However I think the x variable values you choose might be too close in the current figure size to show all the vertical lines (simply maximize the figure to see what I mean)...
To get a better idea of what each axis contains, here is a divided view where the plot on the left contains only the graphics rendered by hAx1, while the plot on right contains only the hAx2 components. Those two views are basically overlayed on top of each other in the final figure shown before.
Unfortunately, while the trick of over- or under-laying a second, gridded axes mostly works, Matlab does not render it properly when you save to a PDF file. This is because Matlab does not support transparency in PDFs.
One workaround is to simply use line to draw in the grid lines one by one:
for dir='XY';
ticks = get(gca, [dir 'Tick']);
lim = get(gca, [dir 'lim']);
for ii=1:length(ticks)
coord = ticks(ii);
for jj=1:9,
if jj==1 % major grid properties
color = [1 1 1]*0.9;
weight = 2;
else % minor grid properties
color = [1 1 1]*0.9;
weight = 1;
end
if jj*coord > lim(2)
continue
end
if dir=='X'
L = line((jj*coord)*[1 1], get(gca, 'ylim'), ...
'color', color, 'linewidth', weight);
else
L = line(get(gca, 'xlim'), (jj*coord)*[1 1], ...
'color', color, 'linewidth', weight);
end
uistack(L, 'bottom');
end
end
end
One downside of this approach is that it overwrites the tick marks and plot boundary box. A solution to this is to combine this approach with the trick of under-laying a second axes. Draw the fake grid on the underlying axes. This IS rendered properly in PDF:
While Amro is right that the minor grid's color is the same as that of the axis labels, you can always turn off the axis labels and overlay a second axes with transparent filling and set the labels on that in a different color. Here's a small example showing how:
plot(rand(10,1))
xTicks=get(gca,'xTick');
yTicks=get(gca,'ytick');
set(gca, 'XMinorGrid','on', 'YMinorGrid','on',...
'XColor','r', 'YColor','g','xticklabel',[],'yticklabel',[],...
'box','off')
h2=axes;
set(h2,'color','none','xtick',linspace(0,1,numel(xTicks)),'xticklabel',xTicks,...
'ytick',linspace(0,1,numel(yTicks)),'yticklabel',yTicks)
This lets you set independent colors for major and minor X and Y grid lines, without overwriting the outer box. Even better, subsequent legend() commands will pick up the plot lines, not the manually drawn grid lines.
The trick is to make copies of the axes, then reverse their order in the figure's drawing hierarchy. Each copy of the axes can then draw its own set of grid colors and styles.
This strategy is compatible with subplot() and print().
function gridcolor(majorX, majorY, minorX, minorY)
ax1 = gca; %# get a handle to first axis
%# create a second transparent axis, same position/extents, same ticks and labels
ax2 = copyobj(ax1,gcf);
ax3 = copyobj(ax1,gcf);
delete(get(ax2,'Children'));
delete(get(ax3,'Children'));
set(ax2, 'Color','none', 'Box','off','YTickLabel',[],'YTickLabel',[],...
'GridLineStyle', '-',...
'XGrid','on','YGrid','on',...
'XMinorGrid','off','YMinorGrid','off',...
'XColor',majorX,'YColor',majorY);
set(ax3,'Box','off','YTickLabel',[],'YTickLabel',[],...
'MinorGridLineStyle','-',...
'XGrid','off','YGrid','off',...
'XMinorGrid','on','YMinorGrid','on',...
'XColor',minorX,'YColor',minorY);
set(ax1, 'Color','none', 'Box','on')
handles = [ax3; ax2; ax1];
c = get(gcf,'Children');
for i=1:length(handles)
c = c(find(c ~= handles(i)));
end
set(gcf,'Children',[c; flipud(handles)]);
linkaxes([ax1 ax2 ax3]);
end
subplot(211);semilogx([1:4000]);gridcolor('r','g','c','b');
subplot(212);semilogx(([1:4000]).^-1);gridcolor('r','g','c','b');