熊猫时间序列图-将鼠标悬停在线上时显示日期

Question

I am plotting a pandas Series where the index is date by date. When I say series.plot(), a chart is generated correctly. The problem is that when I hover the mouse over interesting points on the chart, it only shows the Month and Year of that point. It does not show the exact date of that point.

Below is a sample of the code. Depending on luck, when I mouse over the line, sometimes I see the exact date displayed on the status bar but sometimes I only see year and month.

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
idx = pd.date_range('2011-1-1', '2015-1-1')
x = pd.Series(np.cumsum(np.random.randn(len(idx))), idx)

df = pd.DataFrame(x)
df.plot()
plt.show()

Is there any way to display the exact date? How does matplotlib control what to display on status bar? I wonder it has something to do with pandas changing the default configuration after some code is called.

Answer 1

When launching your code everything seems to be working and a complete date (the x-coordinate) is shown in the status bar all the time. But the two coordinates are shown also when I am not directly over the graph (so it is difficult to know the actual values of your graph). Are you looking for a tooltip that shows the exact date, when mousing over the graph, or are the right values (complete dates) in the status bar enough? Can you make a screenshot of how your problem looks like, when it occurs and provide details on the versions you are using? I am on matplotlib 1.4.3 and numpy 1.9.2 combined with pandas 0.15.2.

Also have a look at the matplotlib recipes ( http://matplotlib.org/users/recipes.html )! Section "Fixing common date annoyances" sounds very similar to your problem!