1. 定义聚合

  1. #  读取flights数据集，查询头部
  2.  In[2]: flights = pd.read_csv('data/flights.csv')
  3.         flights.head()
  4. Out[2]:

1. 定义聚合 - 图1

  1. #  按照AIRLINE分组，使用agg方法，传入要聚合的列和聚合函数
  2.  In[3]: flights.groupby('AIRLINE').agg({'ARR_DELAY':'mean'}).head()
  3. Out[3]:
  1. #  或者要选取的列使用索引，聚合函数作为字符串传入agg
  2.  In[4]: flights.groupby('AIRLINE')['ARR_DELAY'].agg('mean').head()
  3. Out[4]: 
  4. AIRLINE
  5. AA     5.542661
  6. AS    -0.833333
  7. B6     8.692593
  8. DL     0.339691
  9. EV     7.034580
  10. Name: ARR_DELAY, dtype: float64

1. 定义聚合 - 图2

  1. #  也可以向agg中传入NumPy的mean函数
  2.  In[5]: flights.groupby('AIRLINE')['ARR_DELAY'].agg(np.mean).head()
  3. Out[5]:

1. 定义聚合 - 图3

  1. #  也可以直接使用mean()函数
  2.  In[6]: flights.groupby('AIRLINE')['ARR_DELAY'].mean().head()
  3. Out[6]:

1. 定义聚合 - 图4

原理

  1. #  groupby方法产生的是一个DataFrameGroupBy对象
  2.  In[7]: grouped = flights.groupby('AIRLINE')
  3.         type(grouped)
  4. Out[7]: pandas.core.groupby.DataFrameGroupBy

更多

  1. #  如果agg接收的不是聚合函数，则会导致异常
  2.  In[8]: flights.groupby('AIRLINE')['ARR_DELAY'].agg(np.sqrt)
  3. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py:842: RuntimeWarning: invalid value encountered in sqrt
  4.   f = lambda x: func(x, *args, **kwargs)
  5. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py:3015: RuntimeWarning: invalid value encountered in sqrt
  6.   output = func(group, *args, **kwargs)
  8. ---------------------------------------------------------------------------
  9. ValueError                                Traceback (most recent call last)
  10. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py in agg_series(self, obj, func)
  11.    2177         try:
  12. -> 2178             return self._aggregate_series_fast(obj, func)
  13.    2179         except Exception:
  15. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py in _aggregate_series_fast(self, obj, func)
  16.    2197                                     dummy)
  17. -> 2198         result, counts = grouper.get_result()
  18.    2199         return result, counts
  20. pandas/_libs/src/reduce.pyx in pandas._libs.lib.SeriesGrouper.get_result (pandas/_libs/lib.c:39105)()
  22. pandas/_libs/src/reduce.pyx in pandas._libs.lib.SeriesGrouper.get_result (pandas/_libs/lib.c:38973)()
  24. pandas/_libs/src/reduce.pyx in pandas._libs.lib._get_result_array (pandas/_libs/lib.c:32039)()
  26. ValueError: function does not reduce
  28. During handling of the above exception, another exception occurred:
  30. ValueError                                Traceback (most recent call last)
  31. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py in aggregate(self, func_or_funcs, *args, **kwargs)
  32.    2882             try:
  33. -> 2883                 return self._python_agg_general(func_or_funcs, *args, **kwargs)
  34.    2884             except Exception:
  36. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py in _python_agg_general(self, func, *args, **kwargs)
  37.     847             try:
  38. --> 848                 result, counts = self.grouper.agg_series(obj, f)
  39.     849                 output[name] = self._try_cast(result, obj, numeric_only=True)
  41. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py in agg_series(self, obj, func)
  42.    2179         except Exception:
  43. -> 2180             return self._aggregate_series_pure_python(obj, func)
  44.    2181 
  46. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py in _aggregate_series_pure_python(self, obj, func)
  47.    2214                         isinstance(res, list)):
  48. -> 2215                     raise ValueError('Function does not reduce')
  49.    2216                 result = np.empty(ngroups, dtype='O')
  51. ValueError: Function does not reduce
  53. During handling of the above exception, another exception occurred:
  55. Exception                                 Traceback (most recent call last)
  56. <ipython-input-8-2bcc9ccfec77> in <module>()
  57. ----> 1 flights.groupby('AIRLINE')['ARR_DELAY'].agg(np.sqrt)
  59. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py in aggregate(self, func_or_funcs, *args, **kwargs)
  60.    2883                 return self._python_agg_general(func_or_funcs, *args, **kwargs)
  61.    2884             except Exception:
  62. -> 2885                 result = self._aggregate_named(func_or_funcs, *args, **kwargs)
  63.    2886 
  64.    2887             index = Index(sorted(result), name=self.grouper.names[0])
  66. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/groupby.py in _aggregate_named(self, func, *args, **kwargs)
  67.    3015             output = func(group, *args, **kwargs)
  68.    3016             if isinstance(output, (Series, Index, np.ndarray)):
  69. -> 3017                 raise Exception('Must produce aggregated value')
  70.    3018             result[name] = self._try_cast(output, group)
  71.    3019 
  73. Exception: Must produce aggregated value