python - 时间序列混合时出现 TypeError : cannot concatenate a non-NDFrame object,

Question

有一个时间序列 ts (dataframe.to_dict())

{'latitude': {Timestamp('2014-10-20 15:21:56.571000'): 48.145553900000003,
  Timestamp('2014-10-20 15:24:00.789000'): 48.145584300000003,
  Timestamp('2014-10-20 15:26:00.911000'): 48.145497599999999,
  Timestamp('2014-10-20 15:33:57.764000'): 48.145548699999999,
  Timestamp('2014-10-20 15:36:45.760000'): 48.145454999999998},
  'longitude': {Timestamp('2014-10-20 15:21:56.571000'): 11.578263,
  Timestamp('2014-10-20 15:24:00.789000'): 11.5783685,
  Timestamp('2014-10-20 15:26:00.911000'): 11.578193499999999,
  Timestamp('2014-10-20 15:33:57.764000'): 11.5782843,
  Timestamp('2014-10-20 15:36:45.760000'): 11.5783164},
  'speed': {Timestamp('2014-10-20 15:21:56.571000'): 0.0,
  Timestamp('2014-10-20 15:24:00.789000'): 0.0,
  Timestamp('2014-10-20 15:26:00.911000'): 0.0,
  Timestamp('2014-10-20 15:33:57.764000'): 0.0,
  Timestamp('2014-10-20 15:36:45.760000'): 0.0}}

和自定义聚合函数(示例)

def my_func(group):
    first_latitude = group['latitude'].sort_index().head(1).values[0]
    last_longitude = group['longitude'].sort_index().tail(1).values[0]
    return first_latitude - last_longitude

想要用自定义函数聚合时间序列10分钟，所以

ts.groupby(pd.TimeGrouper(freq='10Min')).apply(my_func)

然后它给了我错误，而不是正确的结果

TypeError: cannot concatenate a non-NDFrame object

这个错误说明了什么？我怎样才能正确地编写代码？非常感谢

Answer 1

我认为你想要agg(聚合)，而不是apply，对于你的每个组，你想要1个返回值:

In [185]:

print ts.groupby(pd.TimeGrouper(freq='10Min')).agg(my_func)
                      latitude  longitude      speed
2014-10-20 15:20:00  36.567360  36.567360  36.567360
2014-10-20 15:30:00  36.567232  36.567232  36.567232