Pandastimedeltas的字符串格式

我注意到Pandas知道如何将timedelta对象格式化为字符串.

I noticed that Pandas knows how to smartly format a timedelta object into a string.

In [1]: df[column][rows].max()
Out[1]: 
0   2 days, 02:08:07
dtype: timedelta64[ns]

当我尝试手动执行此操作时，我会不断获取字符串(以纳秒为单位).

When I try to do this manually I keep getting the string in nanoseconds.

In [2]: df[column][rows].max()[0]
Out[2]: numpy.timedelta64(180487000000000,'ns')

In [2]: str(df[column][rows].max()[0])
Out[2]: '180487000000000 nanoseconds'

我不想重蹈覆辙，所以有什么方法可以访问熊猫用来在x days, hh:mm:ss中显示timedelta对象的字符串格式化方法(或字符串本身)?

I would rather not reinvent the wheel, so is there any way to access the string formatting method (or the string itself) that Pandas uses to show a timedelta object in x days, hh:mm:ss ?