pandas.DataFrame(a)
Out[41]:
1 2 3
0 1 2 NaN
1 1 NaN 3
a = [{1:1.0,2:2.0}, {1:1.0,3:3.0}]
pandas.DataFrame(a)
Out[43]:
1 2 3
0 1 2 NaN
1 1 NaN 3
pandas.SparseDataFrame(a)
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-44-50288c1b1994> in <module>()
----> 1 pandas.SparseDataFrame(a)
/Library/Python/2.7/site-packages/pandas/sparse/frame.pyc in __init__(self, data, index, columns, default_kind, default_fill_value)
94 sdict, columns, index = self._init_dict(data, index, columns)
95 elif isinstance(data, (np.ndarray, list)):
---> 96 sdict, columns, index = self._init_matrix(data, index, columns)
97 elif isinstance(data, DataFrame):
98 sdict, columns, index = self._init_dict(data, data.index,
/Library/Python/2.7/site-packages/pandas/sparse/frame.pyc in _init_matrix(self, data, index, columns, dtype)
203
204 data = dict([(idx, data[:, i]) for i, idx in enumerate(columns)])
--> 205 return self._init_dict(data, index, columns, dtype)
206
207 def __array_wrap__(self, result):
/Library/Python/2.7/site-packages/pandas/sparse/frame.pyc in _init_dict(self, data, index, columns, dtype)
174 v = [v.get(i, nan) for i in index]
175
--> 176 v = sp_maker(v)
177 sdict[k] = v
178
/Library/Python/2.7/site-packages/pandas/sparse/frame.pyc in <lambda>(x)
159 kind=self.default_kind,
160 fill_value=self.default_fill_value,
--> 161 copy=True)
162
163 sdict = {}
/Library/Python/2.7/site-packages/pandas/sparse/series.pyc in __new__(cls, data, index, sparse_index, kind, fill_value, name, copy)
127 if sparse_index is None:
128 values, sparse_index = make_sparse(data, kind=kind,
--> 129 fill_value=fill_value)
130 else:
131 values = data
/Library/Python/2.7/site-packages/pandas/sparse/array.pyc in make_sparse(arr, kind, fill_value)
426
427 if np.isnan(fill_value):
--> 428 mask = -np.isnan(arr)
429 else:
430 mask = arr != fill_value
TypeError: ufunc 'isnan' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule 'safe'
最佳答案
Pandas >= 1.0SparseDataFrame和 SparseSeries自 Extension Types 起从 API 中删除现在 Pandas 支持。
使用 SparseArray 扩展类型改为声明稀疏列。
以前的方式
pd.SparseDataFrame({"A": [0, 1]})
新方式 [✓]
pd.DataFrame({"A": pd.arrays.SparseArray([0, 1])})
A
0 0
1 1
关于python - 如何在 Pandas 中创建 SparseDataFrame？，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/25358381/