如何在熊猫中的两个timedelta变量之间绘制图形？

Question

我有以下数据集：

Duration1                            Duration2
05:13:45                             01:09:58
18:53:38                             01:53:18
 NaT                                 01:03:38
07:19:38                             01:23:26

我想在duration1和duration2之间绘制图表？

df['duration1'] =[" 05:13:45 "," 18:53:38 "," NaT ","07:19:38"]
df['duration2'] = [" 01:09:58","01:53:18","01:03:38","01:23:26"]

持续时间1和持续时间2的数据类型是timedelta64 [ns]

额外奖励：是否可以根据绘制的图表趋势获得函数？

Answer 1

使用dt.total_seconds

df.stack().dt.total_seconds().unstack().plot.scatter(
    'Duration1', 'Duration2')

获取趋势线的最简单方法是使用seaborn.regplot

import seaborn as sns

d = df.stack().dt.total_seconds().unstack()
sns.regplot(d.Duration1, d.Duration2, ci=None)

代码从头到尾
你应该可以复制/粘贴

from io import StringIO
import pandas as pd
import seaborn as sns


txt = """Duration1                            Duration2
-1 days +05:13:45                   0 days 01:09:58
-6 days +18:53:38                   0 days 01:53:18
 NaT                                0 days 01:03:38
10 days +07:19:38                   0 days 01:23:26
"""

df = pd.read_csv(StringIO(txt), sep='\s{2,}', engine='python').apply(pd.to_timedelta)

d = df.stack().dt.total_seconds().unstack()

sns.regplot(d.Duration1, d.Duration2, ci=None)