如何在matlab中修复此boxplot的图例？

Question

此箱图的图例颜色相同！我怎样才能解决这个问题？如何将x轴标签移动一点点？谢谢你的帮助。

close all
clc;clear;
f=figure;
Temp_O=[-0.234115422389688;-0.153751688636750;3.03158128172032;-0.746185319551222;0.491616009046725;1.17490826218458;0.495331079652895;0.757394580248284;1.28467417069223;0.710444835069366;-0.979521722186138;-0.216850422633648;0.0596632891728577;-0.525362330358090;0.681608181821661;-0.995216710339821;-0.706416688978551;-0.147700048468633;-0.145946504735073;0.355209739265580;1.25860455564176;0.970569089382961;3.99404165520844;0.433235373567272;1.37023527554759;1.45032207715449;2.00968917969203;0.840884198707613;2.08558564237223;2.05435556980046;-15.5517060656394;3.18991806590028;1.28277879106186;2.15931490153483;3.19647581545030;2.97877640768595;0.0857405478541730;-1.59362648933500;-2.18109410889313;0.751077088333943;0.795072796032814;4.18896005388773;-0.591461781602054;-0.229818549439720];
position_O = 5:5:25;
position_O=position_O';
g = [ones(10,1); 2*ones(10,1); 3*ones(10,1) ;4*ones(10,1);5*ones(4,1)];
box_O = boxplot(Temp_O,g,'colors','b','positions',position_O,'width',0.8);
 h=findobj(gca,'tag','Outliers');
delete(h)
  set(gca,'XTickLabel',{' '}) 
hold on  
Temp_S=[-0.234069549668875;-0.0803021151079149;0.166729084507040;-0.991371043478263;0.320651878289472;0.118699258741257;-0.190944834558825;0.540367970198674;1.02556298920863;0.112849364285713;-0.395341229166667;0.382362326388889;-1.40591456976744;0.247202120000001;-1.33262568333333;-1.27793610544218;0.0400995141843974;-1.32333150653595;-1.84221947163121;0.407607340136054;0.264276120300749;-0.337747273809525;1.03841878571429;-1.41048786507936;0.901727821428570;-1.03012908482143;2.69786876785714;-0.691010535714286;1.66913088345865;0.684260974489794;-10.3923539047619;1.04994314285714;2.13557031632653;3.87736348701299;7.38705700000000;0.0451628482142860;-3.69094742857143;-1.14071104081633;-3.15830153968254;-4.41399970408163;6.09908001655629;0.0267684861111112;-2.67854298170732;0.925146217948717;];
position_S = 6.8:5:26.8;
position_S=position_S';
box_S = boxplot(Temp_S,g,'colors','r','positions',position_S,'width',0.8);
h=findobj(gca,'tag','Outliers');
delete(h) 
legend(findobj(gca,'Tag','Box'),'Group1','Group2')
set(gca,'XTickLabel',{' '}) ;
hold off 
text('Position',[5,-11],'String','S')
text('Position',[10,-11],'String','M')
text('Position',[15,-11],'String','L')
text('Position',[20,-11],'String','V')
text('Position',[25,-11],'String','C')

 xlabel('Types','FontSize',10);
% set(get(gca, 'XLabel'), 'Position', [0 .2 0]); %
ylim([-10.5 7.8]);

Answer 1

问题是你只显示前两个框的图例（你的共有10个框），而且这两个框都是红色的。找到的前5个框是红色，后5个是蓝色。相反，您可以使用第一个和最后一个框。

%// Create the box plot
box_S = boxplot(Temp_S, g, 'colors', 'r', 'positions', position_S, 'width', 0.8);

%// Get all of the box plot objects
boxes = findobj(gca, 'Tag', 'Box');
legend(boxes([end 1]), 'Group1', 'Group2')

你可以通过以下方式更有力地做到这一点：

boxes = findobj(gca, 'Tag', 'box');

%// Sort by xposition
[~, ind] = sort(cellfun(@mean, get(boxes, 'XData')));

%// Apply legends to one red and one blue one.
legend(boxes(ind(1:2)), 'Group1', 'Group2');

要将xlabel移动一点，你可以简单地调整它的Position属性。

yrange = diff(get(gca, 'YLim'));
XL = get(gca, 'XLabel');
original = get(XL, 'Position');

%// Add an extra 1% padding
set(XL, 'Position', original - [0 0.01*yrange 0])

Answer 2

这是一个小例子（我只保留必要的东西）：

x1 = randn(44,1);
x2 = randn(44,1);
pos1 = (5:5:25)';
pos2 = (6.8:5:26.8)';
g = repelem([1 2 3 4 5], [10 10 10 10 4]);

h1 = boxplot(x1, g, 'Colors','b', 'Positions',pos1, 'Width',0.8);
hold on
h2 = boxplot(x2, g, 'Colors','r', 'Positions',pos2, 'Width',0.8);
hold off
legend([h1(5,1),h2(5,1)], {'Group1','Group2'})

引用help boxplot：

%   H = BOXPLOT(...) returns the handle H to the lines in the box plot.
%   H has one column per box, consisting of the handles for the various
%   parts of the box.  For the traditional plotstyle, the rows correspond
%   to: upper whisker, lower whisker, upper adjacent value, lower adjacent
%   value, box, median, and outliers. For the compact plotstyle, the rows
%   correspond to: whiskers, box, median outer, median inner, and outliers.
%   If median comparison intervals are indicated with markers, H will have
%   two more rows for notch lo and notch hi.  If medianstyle or boxstyle 
%   have been set explicitly, the meaning of the rows will adjust 
%   accordingly.