python图分组条形图

我有一个3列数据，如下所示

 clm1                       clm2     clm3
 |["shared","connect"]       13297  |aaaa|
 |["stopped","failed]        25002  |aaaa|
 |["success","obtained"]     11189  |aaaa|
 |["shared","connect"]       16770  |bbbb|
 |["stopped","failed]        81777  |bbbb|
 |["success","obtained"]     9555   |bbbb|

我想在python中使用下面的条形图，我能够编写简单的图形，但无法编写能够对clm3进行分组和绘图的逻辑

---

这里的主要问题是matplotlib认为所有分类数据“A”都代表相同的类别，因此它将它们绘制在“A”的相同位置。我们必须发明一个额外的类别来区分所有这些“A”值。例如，我们可以使用

cumcount（）

对所有值“A”从0到n进行编号。例如：

from matplotlib import pyplot as plt
import pandas as pd

#create toy dataframe
#this part you should have included in your question
#as a Minimal, Complete, and Verifiable example
np.random.seed(1234)
df = pd.DataFrame({"cat": ["A", "B", "C", "C", "B", "C", "A"], "val": np.random.randint(1, 100, 7)})

#add column for multiple cat values and rearrange dataframe
df["cols"] = df.groupby("cat").cumcount()
df1 = df.pivot(index = "cat", columns = "cols", values = "val")
print(df1)

#plot this table
df1.plot.bar(color = "blue", edgecolor = "white")
plt.legend().set_visible(False)
plt.xticks(rotation = 0)
plt.show()

示例数据帧：

cols     0     1     2
cat                   
A     48.0  16.0   NaN
B     84.0  77.0   NaN
C     39.0  54.0  25.0

示例图：

编辑： 我刚刚注意到，在您的情况下，这更容易，因为，尽管您的问题中从未提到过这一点，但您可能希望将其作为类别“clm1”。因此，您可以简化此过程：

from matplotlib import pyplot as plt
import pandas as pd

#create toy dataframe
np.random.seed(1234)
df = pd.DataFrame({"clm1": ["X", "Y", "Z", "X", "Y", "Z"], "clm2": np.random.randint(1, 100, 6), "clm3": ["A", "A", "A", "B", "B", "B"]})

#rearrange dataframe and plot
df.pivot(index = "clm3", columns = "clm1", values = "clm2").plot.bar(edgecolor = "white")
plt.xticks(rotation = 0)
plt.show()

样本输出：