MATLAB柱状图中如何将零值对应图例以线条形式展示？

现有如下向量数据，用于绘制柱状图：

% 当Phi/Psi = +/-10时
CoverageArea_mean_10 = [84.4735,21.1779,6.4247,2.1416]; 
CoverageArea_min_10 = [98.5128,21.1779,6.9007,2.1416]; 
CoverageArea_max_10 = [70.1963,19.0363,5.9488,2.1416];
% 当Phi/Psi = +/-40时
CoverageArea_mean_40 = [0,4.5211,2.3795,0]; 
CoverageArea_min_40 = [92.5640,21.1779,6.9007,2.1416]; 
CoverageArea_max_40 = [0,0.4759,0.2380,0]; 

当前绘图效果：

柱状图中，当Phi/Psi=+/-40时，mean和max类别的第1、4组数值为0，对应位置无柱状条显示，但图例中所有类别仍以柱状样式展示，导致零值项的图例与实际绘图表现不符。

预期效果：

零值对应的类别在图例中以线条形式呈现，而非柱状样式，匹配实际绘图中无柱状条的表现。

现有绘图代码：

%% 绘制覆盖面积柱状图
x = [15,30,45,60];
figure
COVERAGE = [CoverageArea_min_10;CoverageArea_mean_10;CoverageArea_max_10];
COVERAGEAREA = [COVERAGE(:,1)';COVERAGE(:,2)';COVERAGE(:,3)';COVERAGE(:,4)'];
bar(x,COVERAGEAREA);
xlabel( '半功率张角, \Phi_1_/_2 (°)' );
ylabel('覆盖面积 (m²)');
BarNames = {'min','mean','max'};
legend(BarNames,'Location','best');
grid on;
figure
COVERAGE1 = [CoverageArea_min_40;CoverageArea_mean_40;CoverageArea_max_40];
COVERAGEAREA1 = [COVERAGE1(:,1)';COVERAGE1(:,2)';COVERAGE1(:,3)';COVERAGE1(:,4)'];
bar(x,COVERAGEAREA1);
xlabel( '半功率张角, \Phi_1_/_2 (°)' );
ylabel('覆盖面积 (m²)');
BarNames = {'min','mean','max'};
legend(BarNames,'Location','best');
grid on;

要实现零值项图例用线条展示，需要手动控制图例的显示样式，具体步骤和代码如下：

修改后的代码 #

%% 绘制覆盖面积柱状图（优化版）
x = [15,30,45,60];
% 处理Phi/Psi=+/-10的图
figure
COVERAGE = [CoverageArea_min_10;CoverageArea_mean_10;CoverageArea_max_10];
COVERAGEAREA = [COVERAGE(:,1)';COVERAGE(:,2)';COVERAGE(:,3)';COVERAGE(:,4)'];
bar_handles = bar(x,COVERAGEAREA); % 保存每个柱状类别的句柄
xlabel( '半功率张角, \Phi_1_/_2 (°)' );
ylabel('覆盖面积 (m²)');
BarNames = {'min','mean','max'};

% 构建自定义图例项
legend_entries = [];
for i = 1:length(bar_handles)
    if all(COVERAGE(i,:) == 0)
        % 全零类别：创建不可见的线条对象用于图例展示
        line_obj = plot(NaN,NaN,'Color',bar_handles(i).FaceColor,'LineStyle','-','LineWidth',2);
        legend_entries = [legend_entries, line_obj];
    else
        % 非零类别：保留原柱状图句柄
        legend_entries = [legend_entries, bar_handles(i)];
    end
end
legend(legend_entries, BarNames,'Location','best');
grid on;

% 处理Phi/Psi=+/-40的图
figure
COVERAGE1 = [CoverageArea_min_40;CoverageArea_mean_40;CoverageArea_max_40];
COVERAGEAREA1 = [COVERAGE1(:,1)';COVERAGE1(:,2)';COVERAGE1(:,3)';COVERAGE1(:,4)'];
bar_handles1 = bar(x,COVERAGEAREA1);
xlabel( '半功率张角, \Phi_1_/_2 (°)' );
ylabel('覆盖面积 (m²)');

% 构建自定义图例项
legend_entries1 = [];
for i = 1:length(bar_handles1)
    if all(COVERAGE1(i,:) == 0)
        line_obj = plot(NaN,NaN,'Color',bar_handles1(i).FaceColor,'LineStyle','-','LineWidth',2);
        legend_entries1 = [legend_entries1, line_obj];
    else
        legend_entries1 = [legend_entries1, bar_handles1(i)];
    end
end
legend(legend_entries1, BarNames,'Location','best');
grid on;

代码说明 #

1. 通过bar_handles保存每个柱状类别的句柄，获取其颜色信息，保证图例样式与原柱状条颜色一致。
2. 遍历每个类别，判断是否为全零值：
   • 全零类别：创建一个用NaN定义的不可见线条，仅用于图例展示，样式设为实线。
   • 非零类别：直接使用原柱状图句柄，保持柱状图例样式。
3. 将混合了柱状句柄和线条句柄的列表传入legend，实现零值项用线条、非零项用柱状的图例效果。

内容的提问来源于stack exchange，提问作者SH_IQ