pandas 在 matplotlib 中对 x 轴进行排序

Sorting the x axis in matplotlib

提问by DIGSUM

Why does this code not plot the x-axis sorted by 'value'?

为什么此代码不绘制按“值”排序的 x 轴？

import pandas as pd
import matplotlib.pyplot as plt

# creating dataframe
df=pd.DataFrame()
df['name'] = [1,2,3]
df['value'] = [4,3,5]

# sorting dataframe
df.sort_values('value', ascending = False, inplace= True)

# plot
plt.scatter(df['value'],df['name'])
plt.show()

回答by andrew_reece

Given your choice of variable names, plus the seeming confusion surrounding the use of scatter plots, it seems like namemay be a categorical variable that you'd like to plot on the x-axis, sorted by value.

考虑到您对变量名称的选择，加上散点图的使用似乎有些混乱，您似乎想name在 x 轴上绘制一个分类变量，按 排序value。

If that's the case, I'd recommend initially plotting with df.indexas the x-axis, and then changing the tick labels to nameentries. Use reset_index()after sort_valuesto get the correct index ordering.

如果是这种情况，我建议最初使用df.indexx 轴进行绘图，然后将刻度标签更改为name条目。使用reset_index()aftersort_values获得正确的索引排序。

Both Pandas and Pyplot should be able to do this without additional modules, but I had some trouble getting tick labels to line up. Instead, I found Seaborn's pointplot()handled this job without any trouble:

Pandas 和 Pyplot 都应该能够在没有额外模块的情况下做到这一点，但是我在让刻度标签对齐时遇到了一些麻烦。相反，我发现 Seabornpointplot()毫不费力地处理了这项工作：

# sort, then reset index
df = df.sort_values('value', ascending = False).reset_index(drop=True)

import seaborn as sns
ax = sns.pointplot(x=df.index, y=df.value)
ax.set_xlabel("Name")
ax.set_ylabel("Value")

# Use name column to label x ticks
_ = ax.set_xticklabels(df.name.astype(str).values)