python - pandas 中的错误：“缓冲区的维数错误（预期为 1，得到 2）”

Question

我看过几个帖子都有同样的错误，但很抱歉，我仍然无法正确解决我的问题。

这是我的 .py 文件中的代码：

def get_sum(self, data, list_of_items):
    # I coded this return line, which worked according to one of the cells of the .ipynb file
    # Returns an integer summing up all of the 1s in a given column (or set of columns)
    return data.iloc[:, list_of_items].all(axis = 'columns').sum()


def get_list(self, data):
    
    basket = [[i] for i in data.columns]
    product_list = []

    for item in basket:
        # I coded these two lines, which I am unable to test due to the error
        if self.get_sum(data, [item]) >= 5:
            product_list.append(item)
    
    return product_list

此单元格将在 .ipynb 中访问它：

basket_list = groceries.get_list(df)
print(basket_list)
# expected to be something like [0, 2]. In the case of the given sample, [1]

错误出现在 .ipynb 单元格的第一行，但这不可能，因为我只更改了 .py 文件（提供了两个文件）。话虽这么说，我不知道错误到底在哪里，因为它没有指向我写的任何东西。如果有人至少可以提示我做错了什么，那将不胜感激。

编辑：这是数据示例：

  0 1 2
0 0 0 1
1 1 1 1
2 1 1 0
3 0 1 0
4 1 1 1
5 1 1 0

这是完整的错误回溯，

ValueError                                Traceback (most recent call last)
<ipython-input-142-6c2c13d12cb0> in <module>()
----> 1 frequent_itemsets = rule_miner.get_frequent_itemsets(syn_df)
      2 print(frequent_itemsets)

/content/rule_miner.py in get_frequent_itemsets(self, data)
    121                 # this class.
    122 
--> 123                 if self.get_support(data, [itemset]) >= self.support_t:
    124                     new_itemsets.append(itemset)
    125 

/content/rule_miner.py in get_support(self, data, itemset)
     28         # function.
     29 
---> 30         return data.iloc[:, itemset].all(axis = 'columns').sum()
     31 
     32     def merge_itemsets(self, itemsets):

/usr/local/lib/python3.7/dist-packages/pandas/core/indexing.py in __getitem__(self, key)
    871                     # AttributeError for IntervalTree get_value
    872                     pass
--> 873             return self._getitem_tuple(key)
    874         else:
    875             # we by definition only have the 0th axis

/usr/local/lib/python3.7/dist-packages/pandas/core/indexing.py in _getitem_tuple(self, tup)
   1447             pass
   1448 
-> 1449         return self._getitem_tuple_same_dim(tup)
   1450 
   1451     def _get_list_axis(self, key, axis: int):

/usr/local/lib/python3.7/dist-packages/pandas/core/indexing.py in _getitem_tuple_same_dim(self, tup)
    748                 continue
    749 
--> 750             retval = getattr(retval, self.name)._getitem_axis(key, axis=i)
    751             # We should never have retval.ndim < self.ndim, as that should
    752             #  be handled by the _getitem_lowerdim call above.

/usr/local/lib/python3.7/dist-packages/pandas/core/indexing.py in _getitem_axis(self, key, axis)
   1485         # a list of integers
   1486         elif is_list_like_indexer(key):
-> 1487             return self._get_list_axis(key, axis=axis)
   1488 
   1489         # a single integer

/usr/local/lib/python3.7/dist-packages/pandas/core/indexing.py in _get_list_axis(self, key, axis)
   1467         """
   1468         try:
-> 1469             return self.obj._take_with_is_copy(key, axis=axis)
   1470         except IndexError as err:
   1471             # re-raise with different error message

/usr/local/lib/python3.7/dist-packages/pandas/core/generic.py in _take_with_is_copy(self, indices, axis)
   3361         See the docstring of `take` for full explanation of the parameters.
   3362         """
-> 3363         result = self.take(indices=indices, axis=axis)
   3364         # Maybe set copy if we didn't actually change the index.
   3365         if not result._get_axis(axis).equals(self._get_axis(axis)):

/usr/local/lib/python3.7/dist-packages/pandas/core/generic.py in take(self, indices, axis, is_copy, **kwargs)
   3349 
   3350         new_data = self._mgr.take(
-> 3351             indices, axis=self._get_block_manager_axis(axis), verify=True
   3352         )
   3353         return self._constructor(new_data).__finalize__(self, method="take")

/usr/local/lib/python3.7/dist-packages/pandas/core/internals/managers.py in take(self, indexer, axis, verify, convert)
   1455         new_labels = self.axes[axis].take(indexer)
   1456         return self.reindex_indexer(
-> 1457             new_axis=new_labels, indexer=indexer, axis=axis, allow_dups=True
   1458         )
   1459 

/usr/local/lib/python3.7/dist-packages/pandas/core/internals/managers.py in reindex_indexer(self, new_axis, indexer, axis, fill_value, allow_dups, copy, consolidate)
   1289 
   1290         if axis == 0:
-> 1291             new_blocks = self._slice_take_blocks_ax0(indexer, fill_value=fill_value)
   1292         else:
   1293             new_blocks = [

/usr/local/lib/python3.7/dist-packages/pandas/core/internals/managers.py in _slice_take_blocks_ax0(self, slice_or_indexer, fill_value, only_slice)
   1369         else:
   1370             blknos = algos.take_1d(
-> 1371                 self.blknos, slobj, fill_value=-1, allow_fill=allow_fill
   1372             )
   1373             blklocs = algos.take_1d(

/usr/local/lib/python3.7/dist-packages/pandas/core/algorithms.py in take_nd(arr, indexer, axis, out, fill_value, allow_fill)
   1735         arr.ndim, arr.dtype, out.dtype, axis=axis, mask_info=mask_info
   1736     )
-> 1737     func(arr, indexer, out, fill_value)
   1738 
   1739     if flip_order:

pandas/_libs/algos_take_helper.pxi in pandas._libs.algos.take_1d_int64_int64()

ValueError: Buffer has wrong number of dimensions (expected 1, got 2)

标签： pythonpandas

Answer 1

不需要复杂的自定义函数，您可以使用 pandas 向量函数：

获取至少有五个 1 的列的名称：

>>> df.columns[df.eq(1).sum().ge(5)]
Index(['1'], dtype='object')

过滤这些列：

df.loc[:, df.eq(1).sum().ge(5)]

输出：

   1
0  0
1  1
2  1
3  1
4  1
5  1

这个怎么运作：

(df.eq(1) # values equal to 1 -> True
   .sum() # count number of True
   .ge(5) # True if sum ≥ 5
)

OP代码修复：

loc用于代替iloc传递给的错误列表get_sum

def get_sum(data, list_of_items):
    # I coded this return line, which worked according to one of the cells of the .ipynb file
    return data.loc[:, list_of_items].all(axis = 'columns').sum()


def get_list(data):
    
    product_list = []

    for item in df.columns:
        # I coded these two lines, which I am unable to test due to the error
        if get_sum(data, [item]) >= 5:
            product_list.append(item)
    
    return product_list