在绘制来自多个数组的数据时，将线条颜色与图例匹配

Question

我在不同的板上有来自多个设备的温度数据，例如，在板1上，我对PCB本身有温度，对于3种不同的FET，对于2和3板也有类似的温度数据。我将数据读取到一个数据中，并希望为每个被测试的板绘制相同颜色的数据，但是对板上的每个设备都有不同的标记。例如，板1的所有测量值都是蓝色的，PCB温度使用标记'+'，FET1使用标记'v‘等等。

我像这样读文件：

for file_name in glob.glob(path+'*.csv'):
    filename[i] = os.path.basename(file_name)
    print(filename[i])
    #x[i]= np.genfromtxt(path+ filename[i], delimiter=',',skip_header=20,usecols=(2,4,6,8))
    x[i]=pd.read_csv(path+filename[i], header=0,usecols=[2,4,6,8], skiprows=12,names=['PCB', 'FET1', 'FET2', 'FET3'])

并创建一个数据数组。

然后，我绘制不同的列：

colors=['r','b','g','c','m']
for i in range(len(filename)):
    #plt.figure()
    plt.plot(sc.decimate(x[i]['PCB'],5),'-+'+colors[i],label="PCB")
    plt.plot(sc.decimate(x[i]['FET1'],5),'-v'+colors[i],label='FET1')
    plt.plot(sc.decimate(x[i]['FET2'],5),'-x'+colors[i],label='FET2')
    plt.plot(sc.decimate(x[i]['FET3'],5),'-o'+colors[i],label='FET3')
    leg=np.append(leg, filename[i][0:7])
    #plt.show()


plt.show()
plt.legend(leg)

​

​

标记正确显示，但当我迭代数据格式时，颜色信息丢失了。我如何绘制数据并对其进行排列，以便图例在每组线条中使用相同的颜色(按索引I分组)？

下面是一些示例数据:文件1：

Name:,Data Instr INSTR 3/5/2020 11:51:59,,,,,,,,,,,,
Owner:,lab1,,,,,,,,,,,,
Comments:,,,,,,,,,,,,,
Acquisition Date:,3/5/2020 11:51,,,,,,,,,,,,
&Instrument:,34970A,Address:,ASRL11::INSTR,Modules:,1,Slot3:,34901A,,,,,,
Total Channels:,4,,,,,,,,,,,,
Channel,Name,Function,Range,Resolution,AdvSettings,Scale,Gain,Offset,Label,Test,Low,High,HWAlarm
316,PCB_CTR,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
317,Q24,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
318,Q25,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
319,Q18,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
Scan  Control:,Start Action:,Immediately,Stop Action:,User Terminated,,,,,,,,,
Scan,Time,316 <PCB_CTR> (C),Alarm 316,317 <Q24> (C),Alarm 317,318 <Q25> (C),Alarm 318,319 <Q18> (C),Alarm 319,,,,
1,3/5/2020 11:51:59:168,30.471,0,29.241,0,29.165,0,33.302,0,,,,
2,3/5/2020 11:52:01:152,32.197,0,30.634,0,30.564,0,34.819,0,,,,
3,3/5/2020 11:52:03:152,33.795,0,32.019,0,31.879,0,36.848,0,,,,
4,3/5/2020 11:52:05:152,35.315,0,33.383,0,33.236,0,38.282,0,,,,
5,3/5/2020 11:52:07:152,36.965,0,34.734,0,34.62,0,39.946,0,,,,
6,3/5/2020 11:52:09:152,38.255,0,36.054,0,35.776,0,41.18,0,,,,
7,3/5/2020 11:52:11:152,39.467,0,37.328,0,37.028,0,42.258,0,,,,

档案2

Name:,Data Instr INSTR 3/5/2020 10:03:21,,,,,,,,,,,,
Owner:,lab1,,,,,,,,,,,,
Comments:,,,,,,,,,,,,,
Acquisition Date:,3/5/2020 10:03,,,,,,,,,,,,
&Instrument:,34970A,Address:,ASRL11::INSTR,Modules:,1,Slot3:,34901A,,,,,,
Total Channels:,4,,,,,,,,,,,,
Channel,Name,Function,Range,Resolution,AdvSettings,Scale,Gain,Offset,Label,Test,Low,High,HWAlarm
316,PCB_CTR,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
317,Q24,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
318,Q25,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
319,Q18,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
Scan  Control:,Start Action:,Immediately,Stop Action:,User Terminated,,,,,,,,,
Scan,Time,316 <PCB_CTR> (C),Alarm 316,317 <Q24> (C),Alarm 317,318 <Q25> (C),Alarm 318,319 <Q18> (C),Alarm 319,,,,
1,3/5/2020 10:03:21:164,46.334,0,43.755,0,45.706,0,49.129,0,,,,
2,3/5/2020 10:03:22:149,46.997,0,44.262,0,46.35,0,49.773,0,,,,
3,3/5/2020 10:03:23:149,47.615,0,44.671,0,46.974,0,50.402,0,,,,
4,3/5/2020 10:03:24:149,48.267,0,45.229,0,47.628,0,50.879,0,,,,
5,3/5/2020 10:03:25:149,48.861,0,45.711,0,48.164,0,51.495,0,,,,
6,3/5/2020 10:03:26:149,49.455,0,46.323,0,48.783,0,51.9,0,,,,
7,3/5/2020 10:03:27:149,50.014,0,46.796,0,49.351,0,52.334,0,,,,

档案3

Name:,Data Instr INSTR 3/5/2020 13:41:06,,,,,,,,,,,,
Owner:,lab1,,,,,,,,,,,,
Comments:,,,,,,,,,,,,,
Acquisition Date:,3/5/2020 13:41,,,,,,,,,,,,
&Instrument:,34970A,Address:,ASRL11::INSTR,Modules:,1,Slot3:,34901A,,,,,,
Total Channels:,4,,,,,,,,,,,,
Channel,Name,Function,Range,Resolution,AdvSettings,Scale,Gain,Offset,Label,Test,Low,High,HWAlarm
316,PCB_CTR,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
317,Q24,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
318,Q25,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
319,Q18,Temp (Type K),None,C,Temp (Type K)#1#0.016#Auto#0.001#C#Internal#0#false,FALSE,1,0,C,High Only,0,105,Alarm 1
Scan  Control:,Start Action:,Immediately,Stop Action:,User Terminated,,,,,,,,,
Scan,Time,316 <PCB_CTR> (C),Alarm 316,317 <Q24> (C),Alarm 317,318 <Q25> (C),Alarm 318,319 <Q18> (C),Alarm 319,,,,
1,3/5/2020 13:41:06:162,28.121,0,26.882,0,28.785,0,31.061,0,,,,
2,3/5/2020 13:41:08:147,30.582,0,27.873,0,30.691,0,33.024,0,,,,
3,3/5/2020 13:41:10:147,31.782,0,28.935,0,32.578,0,34.876,0,,,,
4,3/5/2020 13:41:12:147,34.003,0,30.094,0,34.247,0,36.652,0,,,,
5,3/5/2020 13:41:14:147,35.097,0,31.199,0,35.975,0,38.142,0,,,,
6,3/5/2020 13:41:16:147,36.708,0,32.334,0,37.504,0,39.721,0,,,,
7,3/5/2020 13:41:18:147,38.274,0,33.508,0,39.048,0,41.198,0,,,,

谢谢你的帮助。

编辑

在@ have 444输入的帮助下，我更接近我想要的东西，但我仍然有问题：

for i in range(len(filename)):
    l=plt.plot(sc.decimate(x[i]['PCB'],5),'-+'+colors[i],label="PCB")
    lines=np.append(lines,l[0].get_label())
    l=plt.plot(sc.decimate(x[i]['FET1'],5),'-v'+colors[i],label='FET1')
    lines=np.append(lines,l[0].get_label())
    l=plt.plot(sc.decimate(x[i]['FET2'],5),'-x'+colors[i],label='FET2')
    lines=np.append(lines,l[0].get_label())
    l=plt.plot(sc.decimate(x[i]['FET3'],5),'-o'+colors[i],label='FET3')
    lines=np.append(lines,l[0].get_label())

    linesclr=np.append(linesclr, l)  # save color info
    names = np.append(names, filename[i][0:7])

fig.legend(lines, loc=1)
fig.legend(linesclr, labels=names, loc=2)
plt.show()

如下所示，我试图添加的第二个图例没有显示正确的颜色，即每个文件读取1种颜色(左上角)：

​

​

我不明白为什么左边的图例没有显示右颜色，因为颜色信息位于lineslr数组中的每个元素中：

linesclr[0].get_color()
Out[4]: 'r'
linesclr[0].get_color()
Out[5]: 'r'
linesclr[1].get_color()
Out[6]: 'b'
linesclr[2].get_color()
Out[7]: 'g'

此外，我不明白为什么标记并不总是那个传说中所有键的圆圈('o')。

我正在寻找的解决方案：

在我看来，传达情节信息的最好方法是用两个图例，但右边的图例只显示每种设备类型的每一行的标记(黑线带有相应的标记)，左边的图例显示文件名和颜色(没有标记)，用于该文件的所有温度读数。

所以，我想让右边的图例显示四个热电偶的位置和它们的标记：

+ PCB
v FET1 
x FET2 
o FET3

左边的图例显示: ACI50#5用红色，ACI50#用蓝色，ACI50#6用绿色，KDE5515用青色(不管我读了多少文件，每个文件都有相应的情节颜色)。

我试过在传说中的素材库上阅读和写作，并在互联网上寻找过例子，但我在理解我所读的东西方面并不成功！

Answer 1

我希望这就是你要找的东西:)

我认为有两个解决方案是可能的：

在此之前，抱歉我不得不模拟数据，这样情节看起来就不像有序的行了，但我认为这是没有问题的。

1. ，具有所有文件的一个图例：

import glob
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import scipy.signal as sc
import numpy as np

dfs = [] # store df
cmap = cm.get_cmap('Set1')
cols = {}

for i, fn in enumerate(glob.glob("*.csv")) :
    #dfs.append(pd.read_csv(fn, header=0, usecols=[2,4,6,8], skiprows=12, names=['PCB', 'FET1', 'FET2', 'FET3'])) # Uncommenting this line to read from your files should work
    dfs.append(pd.DataFrame(np.random.randn(100, 4), columns=['PCB', 'FET1', 'FET2', 'FET3'])) # Just random data
    cols[i] = cmap(i) # Maps one color to one file with a dict

mrks = {"PCB":'+',"FET1":'v',"FET2":'x',"FET3":'o'} # Maps one sensor to one marker type

fig, ax = plt.subplots(figsize=(12,12))
for n, d in enumerate(dfs) :
    ax.plot(sc.decimate(d['PCB'],5), ls='-', marker=mrks['PCB'], color=cols[n], label="PCB") # Use label to map to files
    ax.plot(sc.decimate(d['FET1'],5), ls='-', marker=mrks['FET1'], color=cols[n], label="FET1")
    ax.plot(sc.decimate(d['FET2'],5), ls='-', marker=mrks['FET2'], color=cols[n], label="FET2")
    ax.plot(sc.decimate(d['FET3'],5), ls='-', marker=mrks['FET3'], color=cols[n], label="FET3")

ax.legend()
plt.show()

给出了这个情节：

​

​

带有分隔传说的

1. (抱歉，在我的示例中，我交换了位置，但您可以轻松地调整它，

import glob
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import scipy.signal as sc
import numpy as np
from matplotlib.lines import Line2D

# create a marker for each thermocouple
mrks = {"PCB":'+',"FET1":'v',"FET2":'x',"FET3":'o'}
marker_legend = [Line2D([0], [0], lw=1, color="k", marker=v, label=k) for k, v in mrks.items()]
color_legend = []

dfs = [] # store df
cmap = cm.get_cmap('Set1')
cols = {}
for i, fn in enumerate(glob.glob("*.csv")) : # read files and map colors to each
    #dfs.append(pd.read_csv(fn, header=0, usecols=[2,4,6,8], skiprows=12, names=['PCB', 'FET1', 'FET2', 'FET3']))
    dfs.append(pd.DataFrame(np.random.randn(100, 4), columns=['PCB', 'FET1', 'FET2', 'FET3']))
    cols[i] = cmap(i)
    color_legend.append(Line2D([0], [0], color=cmap(i), lw=1, label=fn))

fig, ax = plt.subplots(figsize=(12,12))
for n, d in enumerate(dfs) :
    ax.plot(sc.decimate(d['PCB'],5), ls='-', marker=mrks['PCB'], color=cols[n])
    ax.plot(sc.decimate(d['FET1'],5), ls='-', marker=mrks['FET1'], color=cols[n])
    ax.plot(sc.decimate(d['FET2'],5), ls='-', marker=mrks['FET2'], color=cols[n])
    ax.plot(sc.decimate(d['FET3'],5), ls='-', marker=mrks['FET3'], color=cols[n])

first_legend = plt.legend(handles=marker_legend, loc="upper left")
ax = plt.gca().add_artist(first_legend)
second_legend = plt.legend(handles=color_legend, loc="upper right")

plt.show()

这就是由此产生的阴谋：

​

​

如果您不想使用matplotlib中的cmap，您仍然可以创建一个颜色列表，您知道这些颜色的长度将超过您将从其中读取和绘制的文件数，而类似这样的内容：

cmap = ["r","g","b","cyan", ...]
...
for i, fn in enumerate(glob.glob("*.csv")) :
   ...
   cols[i] = cmap[i]
   ...

Answer 2

循环中由同一调用绘制的每个绘图都重写了图例，尽管数据不同。因此，您需要保存它们，然后在不可能进行覆盖之后将它们添加到图例中。

fig = plt.figure()
lines = []
colors=['r','b','g','c','m']
for i in range(len(filename)):
    l=plt.plot(sc.decimate(x[i]['PCB'],5),'-+'+colors[i],label="PCB")
    lines=np.append(lines,l)
    l=plt.plot(sc.decimate(x[i]['FET1'],5),'-v'+colors[i],label='FET1')
    lines=np.append(lines,l)
    l=plt.plot(sc.decimate(x[i]['FET2'],5),'-x'+colors[i],label='FET2')
    lines=np.append(lines,l)
    l=plt.plot(sc.decimate(x[i]['FET3'],5),'-o'+colors[i],label='FET3')
    lines=np.append(lines,l)

fig.legend(lines)
plt.show()

这将提供一个具有交替颜色的len(filename)*4大小图例。如果只想给出每个文件的颜色，可以在每个i中保存一行。

fig = plt.figure()
lines = []
names = []
colors=['r','b','g','c','m']
for i in range(len(filename)):
    l=plt.plot(sc.decimate(x[i]['PCB'],5),'-+'+colors[i],label="PCB")
    lines=np.append(lines,l)
    names=np.append(names, filename[i][0:7])
    plt.plot(sc.decimate(x[i]['FET1'],5),'-v'+colors[i],label='FET1')
    plt.plot(sc.decimate(x[i]['FET2'],5),'-x'+colors[i],label='FET2')
    plt.plot(sc.decimate(x[i]['FET3'],5),'-o'+colors[i],label='FET3')

fig.legend(lines,labels=names)
plt.show()

编辑：作为第三个选项，您还可以直接为第一行集合存储图例句柄，并在以后设置它们。要使其正确工作，您需要将其他人设置为没有传奇。假设你没有在你的情节中显示这些标签，它也是一个选项。

lines = []
names = []
colors=['r','b','g','c','m']
for i in range(len(filename)):
    plt.plot(sc.decimate(x[i]['PCB'],5),'-+'+colors[i],label="PCB")
    handles, labels = plt.gca().get_legend_handles_labels()
    lines=np.append(lines,handles[0])
    names=np.append(names, filename[i][0:7])
    plt.plot(sc.decimate(x[i]['FET1'],5),'-v'+colors[i],label='_nolegend_')
    plt.plot(sc.decimate(x[i]['FET2'],5),'-x'+colors[i],label='_nolegend_')
    plt.plot(sc.decimate(x[i]['FET3'],5),'-o'+colors[i],label='_nolegend_')

plt.legend(lines,labels=names)
plt.show()