5. 在DataFrame上使用运算符

  1. #  college数据集的值既有数值也有对象，整数5不能与字符串相加
  2.  In[37]: college = pd.read_csv('data/college.csv')
  3.          college + 5
  4. ---------------------------------------------------------------------------
  5. TypeError                                 Traceback (most recent call last)
  6. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/ops.py in na_op(x, y)
  7.    1175             result = expressions.evaluate(op, str_rep, x, y,
  8. -> 1176                                           raise_on_error=True, **eval_kwargs)
  9.    1177         except TypeError:
  11. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/computation/expressions.py in evaluate(op, op_str, a, b, raise_on_error, use_numexpr, **eval_kwargs)
  12.     210         return _evaluate(op, op_str, a, b, raise_on_error=raise_on_error,
  13. --> 211                          **eval_kwargs)
  14.     212     return _evaluate_standard(op, op_str, a, b, raise_on_error=raise_on_error)
  16. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/computation/expressions.py in _evaluate_numexpr(op, op_str, a, b, raise_on_error, truediv, reversed, **eval_kwargs)
  17.     121     if result is None:
  18. --> 122         result = _evaluate_standard(op, op_str, a, b, raise_on_error)
  19.     123 
  21. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/computation/expressions.py in _evaluate_standard(op, op_str, a, b, raise_on_error, **eval_kwargs)
  22.      63     with np.errstate(all='ignore'):
  23. ---> 64         return op(a, b)
  24.      65 
  26. TypeError: must be str, not int
  28. During handling of the above exception, another exception occurred:
  30. TypeError                                 Traceback (most recent call last)
  31. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in eval(self, func, other, raise_on_error, try_cast, mgr)
  32.    1183             with np.errstate(all='ignore'):
  33. -> 1184                 result = get_result(other)
  34.    1185 
  36. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in get_result(other)
  37.    1152             else:
  38. -> 1153                 result = func(values, other)
  39.    1154 
  41. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/ops.py in na_op(x, y)
  42.    1201                     with np.errstate(all='ignore'):
  43. -> 1202                         result[mask] = op(xrav, y)
  44.    1203             else:
  46. TypeError: must be str, not int
  48. During handling of the above exception, another exception occurred:
  50. TypeError                                 Traceback (most recent call last)
  51. <ipython-input-37-4749f68a2501> in <module>()
  52.       1 college = pd.read_csv('data/college.csv')
  53. ----> 2 college + 5
  55. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/ops.py in f(self, other, axis, level, fill_value)
  56.    1239                 self = self.fillna(fill_value)
  57.    1240 
  58. -> 1241             return self._combine_const(other, na_op)
  59.    1242 
  60.    1243     f.__name__ = name
  62. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/frame.py in _combine_const(self, other, func, raise_on_error)
  63.    3541     def _combine_const(self, other, func, raise_on_error=True):
  64.    3542         new_data = self._data.eval(func=func, other=other,
  65. -> 3543                                    raise_on_error=raise_on_error)
  66.    3544         return self._constructor(new_data)
  67.    3545 
  69. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in eval(self, **kwargs)
  70.    3195 
  71.    3196     def eval(self, **kwargs):
  72. -> 3197         return self.apply('eval', **kwargs)
  73.    3198 
  74.    3199     def quantile(self, **kwargs):
  76. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in apply(self, f, axes, filter, do_integrity_check, consolidate, **kwargs)
  77.    3089 
  78.    3090             kwargs['mgr'] = self
  79. -> 3091             applied = getattr(b, f)(**kwargs)
  80.    3092             result_blocks = _extend_blocks(applied, result_blocks)
  81.    3093 
  83. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in eval(self, func, other, raise_on_error, try_cast, mgr)
  84.    1189             raise
  85.    1190         except Exception as detail:
  86. -> 1191             result = handle_error()
  87.    1192 
  88.    1193         #  technically a broadcast error in numpy can 'work' by returning a
  90. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in handle_error()
  91.    1172                 #  The 'detail' variable is defined in outer scope.
  92.    1173                 raise TypeError('Could not operate %s with block values %s' %
  93. -> 1174                                 (repr(other), str(detail)))  #  noqa
  94.    1175             else:
  95.    1176                 #  return the values
  97. TypeError: Could not operate 5 with block values must be str, not int
  1. #  行索引名设为INSTNM，用UGDS_过滤出本科生的种族比例
  2. In[38]: college = pd.read_csv('data/college.csv', index_col='INSTNM')
  3.         college_ugds_ = college.filter(like='UGDS_')
  4. In[39]: college == 'asdf' #  这是jn上的，想要比较college和‘asdf’，没有意义，忽略
  5. ---------------------------------------------------------------------------
  6. TypeError                                 Traceback (most recent call last)
  7. <ipython-input-39-697c8af60bcf> in <module>()
  8. ----> 1 college == 'asdf'
  10. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/ops.py in f(self, other)
  11.    1302             #  straight boolean comparisions we want to allow all columns
  12.    1303             #  (regardless of dtype to pass thru) See # 4537 for discussion.
  13. -> 1304             res = self._combine_const(other, func, raise_on_error=False)
  14.    1305             return res.fillna(True).astype(bool)
  15.    1306 
  17. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/frame.py in _combine_const(self, other, func, raise_on_error)
  18.    3541     def _combine_const(self, other, func, raise_on_error=True):
  19.    3542         new_data = self._data.eval(func=func, other=other,
  20. -> 3543                                    raise_on_error=raise_on_error)
  21.    3544         return self._constructor(new_data)
  22.    3545 
  24. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in eval(self, **kwargs)
  25.    3195 
  26.    3196     def eval(self, **kwargs):
  27. -> 3197         return self.apply('eval', **kwargs)
  28.    3198 
  29.    3199     def quantile(self, **kwargs):
  31. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in apply(self, f, axes, filter, do_integrity_check, consolidate, **kwargs)
  32.    3089 
  33.    3090             kwargs['mgr'] = self
  34. -> 3091             applied = getattr(b, f)(**kwargs)
  35.    3092             result_blocks = _extend_blocks(applied, result_blocks)
  36.    3093 
  38. /Users/Ted/anaconda/lib/python3.6/site-packages/pandas/core/internals.py in eval(self, func, other, raise_on_error, try_cast, mgr)
  39.    1203 
  40.    1204                 raise TypeError('Could not compare [%s] with block values' %
  41. -> 1205                                 repr(other))
  42.    1206 
  43.    1207         #  transpose if needed
  45. TypeError: Could not compare ['asdf'] with block values
  1. #  查看前5行
  2.  In[40]: college_ugds_.head()
  3. Out[40]:

5. 在DataFrame上使用运算符 - 图1

  1. #  现在都是均质数据了，可以进行数值运算
  2.  In[41]: college_ugds_.head() + .00501
  3. Out[41]:

5. 在DataFrame上使用运算符 - 图2

  1. #  用底除计算百分比分数
  2.  In[42]: (college_ugds_.head() + .00501) // .01
  3. Out[42]:

5. 在DataFrame上使用运算符 - 图3

  1. #  再除以100
  2.  In[43]: college_ugds_op_round = (college_ugds_ + .00501) // .01 / 100
  3.          college_ugds_op_round.head()
  4. Out[43]:

5. 在DataFrame上使用运算符 - 图4

  1. #  保留两位小数
  2.  In[44]: college_ugds_round = (college_ugds_ + .00001).round(2)
  3.          college_ugds_round.head()
  4. Out[44]:

5. 在DataFrame上使用运算符 - 图5

  1.  In[45]: .045 + .005
  2. Out[45]: 0.049999999999999996
  1.  In[46]: college_ugds_op_round.equals(college_ugds_round)
  2. Out[46]: True

更多

  1. #  DataFrame的通用函数也可以实现上述方法
  2.  In[47]: college_ugds_op_round_methods = college_ugds_.add(.00501).floordiv(.01).div(100)