我有一个由一些 non-unique 非连续的id号索引的pandas数据帧 .

x class'pandas.core.frame.DataFrame'Int64Index:814061条目,8536896到8498857数据列(共1列):收到814061非空值dtypes:datetime64ns

x ['received']是不同长度的时间戳 .

x.ix[i] might have len() == 20  
x.ix[j] might have len() == 32.

对于任何x.ix [i]我可以将时间戳放入[0,1]范围内

df['totalseconds'] = x.ix[i]['received']-x.ix[i]['received'].min()
y = x.ix[i]['received'].max()-x.ix[i]['received'].min()
z = timedelta.total_seconds(y)
df['seconds'] = df['totalseconds'].apply(lambda x: x / timedelta64(1, 's'))
df['norm'] = df['seconds']/z

我正在尝试将x中每个索引的时间段标准化 . 但由于某种原因我遇到了麻烦 .

tldr;如何通过索引ID将所有时间戳记放入[0,1]范围?