Matplotlib子图图例样式不匹配问题排查与解决

我用Matplotlib绘制子图对比Testdata1和Testdata2中的多组参数，为统一管理图例，采用fig.legend()在图层面添加全局图例，而非给每个子图单独设置。

代码可正常运行，但偶尔出现图例样式错乱：Dataset2本该显示黑色折线+方形标记，却变成红色折线+圆形标记。尝试用axs[row, col].get_legend_handles_labels()提取图例句柄，返回空值。

排查后确认根源：输入的测量文件存在重名参数，当第一个子图绘制的是重名参数时，图例就会异常；若第一个子图使用唯一参数，图例显示正常。目前临时方案是确保第一个子图用唯一参数，但需要更可靠的解决办法。

原始代码 #

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import math

paths = ['Testdata1.xlsx', 'Testdata2.xlsx']
Sheets = [['List1'],['List2']]

# define the labels of the plot
Labels = ['Dataset1','Dataset2']
Stil =      ['ro-',     'ks-' ] 

# define the x-axis parameter, label and range
x_axis ="P_EFF_ME" # Normname  Samme x-akse alle plot
x_axis_label = "x-axes title"  

# define the parameters to plot
ParamTitles =   ["A-heading",              "B-heading"   ,    "C-heading"    ,          "D-heading"             ] # Plot heading
ParamLabels =   ["A"             ,        " B",               " C",                     " D"                    ] # Y-axis text
ParamSelect =   ["A_data",                "B_data"    ,       "C_data"  ,               "D_data"                ] # Parametername

# calculate the number of rows and columns for the subplots
num_plots = len(ParamSelect) # number of plots is the same as numbers of elements in ParamSelect
num_cols = min(num_plots, 2) # number of columns is defined as the smallest of either num_plots or 2. If num_plots = 3, then num_cols will still be 2.
num_rows = math.ceil(num_plots / num_cols) # number of rows is found by rounding up num_plots/num_cols. If num_plots = 5 and num_cols = 2, then num_plots/num_cols = 2,5. math.ceil rounds up to 3.

fig, axs = plt.subplots(num_rows, num_cols, figsize=(9, 3 * num_rows))#,layout='constrained') 
fig.subplots_adjust(hspace=0.5,wspace=0.3,top=0.9,bottom=0.075) # adjusts margins

for k, path in enumerate(paths):
    for j, sheet in enumerate(Sheets[k]):
        # read in the data for the current sheet and arranges it as desired for further plotting
        df = pd.read_excel(path, sheet_name=sheet).T             
        df.index=np.arange(0,df.shape[0],1) 
        df.drop([0,2,3], axis=0,inplace=True) 
        df.columns=df.iloc[0,:]      
        df.index=np.arange(0,df.shape[0],1)         
        df.drop([0], axis=0,inplace=True) 
        
        for i, parameter in enumerate(ParamSelect):
            # determine the row and column index for the subplot
            row = i // num_cols
            col = i % num_cols
            # plot the data on the appropriate subplot and add a title
            axs[row, col].plot(df[x_axis], df[parameter],Stil[k])
            axs[row, col].set_xlabel(x_axis_label)
            axs[row, col].set_ylabel(ParamLabels[i])
            axs[row, col].set_title(ParamTitles[i]) # add title
                
fig.legend(labels=Labels, loc='outside upper center', bbox_to_anchor=(0.5, 1), ncol=len(Labels), frameon=False)

plt.show()

复现问题的简化代码 #

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import math

paths = ['Testdata1', 'Testdata2']
Sheets = [['List1'],['List2']]

# define the labels of the plot
Labels = ['Dataset1','Dataset2']
Stil =      ['ro-',     'ks-']

Testdata1=pd.DataFrame({"B0_data":["B_data",3,3,4,1],"C0_data":["C_data",5,3,6,8],"A1_data":["A_data",1,3,5,7],"D0_data":["D_data",9,5,6,4],"":[0,0,0,0,0],"A0_data":["A_data",1.1,3.1,5.1,7.1],"P0_EFF_ME":["P_EFF_ME",0,10,20,30]})
Testdata2=pd.DataFrame({"B0_data":["B_data",2.4,2.4,3.2,0.8],"C0_data":["C_data",4,2.4,4.8,6.4],"A1_data":["A_data",0.8,2.4,4,5.6],"D0_data":["D_data",7.2,4,4.8,3.2],"A0_data":["A_data",0.8,2.4,4,5.6],"P0_EFF_ME":["P_EFF_ME",0,8,16,24]})

# define the x-axis parameter, label and range
x_axis ="P_EFF_ME" # Normname  Samme x-akse all plots
x_axis_label = "x-axes title"  

# define the parameters to plot|
ParamTitles =   ["A-heading", "B-heading" , "C-heading" ,  "D-heading" ] # Plot heading
ParamLabels =   ["A"   ,      " B",         " C",          " D"        ] # Y-axis text
ParamSelect =   ["A_data",    "B_data" ,    "C_data"  ,    "D_data"    ] # Parametername
# If for example B_data is used first, the legends gets correct
#ParamSelect =   ["B_data",    "A_data" ,    "C_data"  ,    "D_data"    ] # Parametername

# calculate the number of rows and columns for the subplots
num_plots = len(ParamSelect) # number of plots is the same as numbers of elements in ParamSelect
num_cols = min(num_plots, 2) # number of columns is defined as the smallest of either num_plots or 2. If num_plots = 3, then num_cols will still be 2.
num_rows = math.ceil(num_plots / num_cols) # number of rows is found by rounding up num_plots/num_cols. If num_plots = 5 and num_cols = 2, then num_plots/num_cols = 2,5. math.ceil rounds up to 3.

fig, axs = plt.subplots(num_rows, num_cols, figsize=(9, 3 * num_rows))#,layout='constrained') 
fig.subplots_adjust(hspace=0.5,wspace=0.3,top=0.9,bottom=0.075) # adjusts margins

for k, path in enumerate(paths):
    for j, sheet in enumerate(Sheets[k]):
        # read in the data 
        df = globals()[path]
        df.index=np.arange(0,df.shape[0],1) 
        df.columns=df.iloc[0,:] # Set column name       
        df.index=np.arange(0,df.shape[0],1)         
        df.drop([0], axis=0,inplace=True) #Sletter opprinnelig Normname kolonne
        for i, parameter in enumerate(ParamSelect):
            # determine the row and column index for the subplot
            row = i // num_cols
            col = i % num_cols
            # plot the data on the appropriate subplot and add a title
            axs[row, col].plot(df[x_axis], df[parameter],Stil[k])
            axs[row, col].set_xlabel(x_axis_label)
            axs[row, col].set_ylabel(ParamLabels[i])
            axs[row, col].set_title(ParamTitles[i]) # add title
                
fig.legend(labels=Labels, loc='outside upper center', bbox_to_anchor=(0.5, 1), ncol=len(Labels), frameon=False)
plt.show()

可靠解决方案 #

方法1：手动创建图例句柄（最直接） #

既然两个数据集的样式是固定的，直接手动创建匹配样式的图例句柄，完全不依赖绘图时的自动收集，彻底规避重名参数的影响。

核心修改代码：

from matplotlib.lines import Line2D

# 在绘图完成后，手动生成图例的handles
handles = []
for style in Stil:
    # 解析样式字符串的颜色、标记、线型
    color = style[0]
    marker = style[1]
    linestyle = style[2] if len(style) > 2 else '-'
    # 创建自定义线条句柄
    handle = Line2D([], [], color=color, marker=marker, linestyle=linestyle, markersize=8)
    handles.append(handle)

# 使用手动创建的handles生成全局图例
fig.legend(handles=handles, labels=Labels, loc='outside upper center', bbox_to_anchor=(0.5, 1), ncol=len(Labels), frameon=False)

方法2：显式指定label并收集首个子图的有效句柄 #

在绘图时给每个数据集的线条显式添加label，仅在第一个子图添加即可（避免重复），然后从第一个子图提取句柄和标签，再生成全局图例。

核心修改代码：

# 修改绘图循环部分
for k, path in enumerate(paths):
    for j, sheet in enumerate(Sheets[k]):
        # ...（数据读取部分不变）
        for i, parameter in enumerate(ParamSelect):
            row = i // num_cols
            col = i % num_cols
            # 仅在第一个子图添加label，其余子图只绘图
            if i == 0:
                axs[row, col].plot(df[x_axis], df[parameter], Stil[k], label=Labels[k])
            else:
                axs[row, col].plot(df[x_axis], df[parameter], Stil[k])
            # ...（子图设置部分不变）

# 从第一个子图提取有效句柄和标签
handles, labels = axs[0, 0].get_legend_handles_labels()
fig.legend(handles=handles, labels=labels, loc='outside upper center', bbox_to_anchor=(0.5, 1), ncol=len(Labels), frameon=False)

方法3：预处理数据合并重名参数 #

如果重名参数是同一指标的重复测量，可以在数据读取阶段合并这些重名列，避免绘图时重复绘制相同参数导致的样式冲突。例如对重名列取均值或保留其中一列：

# 在数据读取后添加重名列合并逻辑
df = df.T
# ...（原有数据处理步骤）
# 按列名分组，取均值合并重名列
df = df.groupby(df.columns, axis=1).mean()

修改后的简化代码（方法1实现） #

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import math
from matplotlib.lines import Line2D

paths = ['Testdata1', 'Testdata2']
Sheets = [['List1'],['List2']]

# define the labels of the plot
Labels = ['Dataset1','Dataset2']
Stil =      ['ro-',     'ks-']

Testdata1=pd.DataFrame({"B0_data":["B_data",3,3,4,1],"C0_data":["C_data",5,3,6,8],"A1_data":["A_data",1,3,5,7],"D0_data":["D_data",9,5,6,4],"":[0,0,0,0,0],"A0_data":["A_data",1.1,3.1,5.1,7.1],"P0_EFF_ME":["P_EFF_ME",0,10,20,30]})
Testdata2=pd.DataFrame({"B0_data":["B_data",2.4,2.4,3.2,0.8],"C0_data":["C_data",4,2.4,4.8,6.4],"A1_data":["A_data",0.8,2.4,4,5.6],"D0_data":["D_data",7.2,4,4.8,3.2],"A0_data":["A_data",0.8,2.4,4,5.6],"P0_EFF_ME":["P_EFF_ME",0,8,16,24]})

# define the x-axis parameter, label and range
x_axis ="P_EFF_ME" # Normname  Samme x-akse all plots
x_axis_label = "x-axes title"  

# define the parameters to plot
ParamTitles =   ["A-heading", "B-heading" , "C-heading" ,  "D-heading" ] # Plot heading
ParamLabels =   ["A"   ,      " B",         " C",          " D"        ] # Y-axis text
ParamSelect =   ["A_data",    "B_data" ,    "C_data"  ,    "D_data"    ] # Parametername

# calculate the number of rows and columns for the subplots
num_plots = len(ParamSelect)
num_cols = min(num_plots, 2)
num_rows = math.ceil(num_plots / num_cols)

fig, axs = plt.subplots(num_rows, num_cols, figsize=(9, 3 * num_rows))
fig.subplots_adjust(hspace=0.5,wspace=0.3,top=0.9,bottom=0.075) # adjusts margins

for k, path in enumerate(paths):
    for j, sheet in enumerate(Sheets[k]):
        # read in the data 
        df = globals()[path]
        df.index=np.arange(0,df.shape[0],1) 
        df.columns=df.iloc[0,:] # Set column name       
        df.index=np.arange(0,df.shape[0],1)         
        df.drop([0], axis=0,inplace=True)
        for i, parameter in enumerate(ParamSelect):
            row = i // num_cols
            col = i % num_cols
            # plot the data on the appropriate subplot
            axs[row, col].plot(df[x_axis], df[parameter],Stil[k])
            axs[row, col].set_xlabel(x_axis_label)
            axs[row, col].set_ylabel(ParamLabels[i])
            axs[row, col].set_title(ParamTitles[i])

# 手动创建图例句柄
handles = []
for style in Stil:
    color = style[0]
    marker = style[1]
    linestyle = style[2] if len(style) > 2 else '-'
    handle = Line2D([], [], color=color, marker=marker, linestyle=linestyle, markersize=8)
    handles.append(handle)

# 添加全局图例
fig.legend(handles=handles, labels=Labels, loc='outside upper center', bbox_to_anchor=(0.5, 1), ncol=len(Labels), frameon=False)

plt.show()

内容的提问来源于stack exchange，提问作者KVa